Unit 13 : Research Methods and

Statistics

Variables
● Something that can be changed or varied
● They help determine if changes to one thing result in changes to another.

Categorical Variables: Variable data points that are placed into qualitatively different
different categories.

Measured variables: data points that appear along a scale and are separated from one
another and where the scale can be discrete or continuous

Descriptive Statistics are summaries of gathered data.

Sample statistics include a measure of:

● Central tendency: Median Mode mean
● Measure of dispersion Standard deviation, variance, mean deviation, variation ration,
range

In experimental studies the independent variable is the categorical variable and the dependent
variable is the measured variable.

Different levels of measurement

1. Nominal Level of Measurement
a. It is quantitative
b. The nominal level is a simple classification. It refers to the data that are categorical.
c. Must be placed in just one category for the purpose of comparisons.
d. We assign numbers to distinguish one item from another. Eg: Maths = 1, Science = 2,
Arts = 3
e. These are separated groups and irreverent differences are ignored for the purpose at
hand

2. Ordinal level of measurement

a. When we put things in order we use ranks, and when we do this we are using ordinal
level of measurement
b. Identifies and describes the magnitude of the variables
c. They rank data by taking the median value
d. Ranks are ordinal level of measurement whereas score are the interval
 e. Ranked without establishing the degree of variation
f. Qualitative

3. Interval level of measurement

a. Interval scale is a more precise measurement
b. It is qualitative in which difference between the variables is meaningful
c. Eg: Temperature: C or F, IQ tests
4. Ratio Level of measurement
a. Qualitative measure
b. Presence if an absolute 0 indicates the absence of a particular variable measure
c. Example: number of items recalled in a memory test, reaction time etc

Difference between Ratio and Interval level of measurement

In statistical analysis differentiating between interval and ratio scale isn’t challenging as it
does not affect the choice of statistical test. To differentiate only think whether zero indicates
the absence of the variable being measured if it does, it is a ratio scale. If it is not, it is an
interval scale.

Types of Variables
1. Independent variable: manipulated by the experimenter
2. Dependent variable: measured and is influenced by the independent variable
3. Extraneous variable: may have an impact on the relationship between the IV and DV
4. Confounding variable: cannot be controlled.
a. Has an impact on the DV
b. Therefore, makes it difficult to determine if the results are due to the influence
of the independent variable, confounding variables or an interaction between
the two.

Parametric and Non-Parametric Tests

● According to the normality of the sample we choose the test
● Sample me not be normally distributed
● When we get a set of data or sample the first thing we need to do is test whether the
sample was normally distributed or not. Based on that we decide what has to be done
or what analysis has to be done.
● To find the correlation between two variables we need to first test the normality of the
sample
● An assessment of the normality of the data is a prerequisite for many statistical tests
because normal data is an underlying assumption in parametric testing.
● The main test of normality are:
○ Kolmogrorov - Smirson test > 2000
○ Shapiro - Wilk Test < 2000
 ● The significance level is the probability of rejecting the null hypothesis when it is
true. For example, a significance level of 0.05 indicates a 5% risk of concluding that a
difference exists when there is no actual difference.
● 0.05 is a confidence level which means there is a 95% confidence level and 5%
chance that this score will not be valid (95% confidence interval)
● The p-value is derived after doing the test. If it is less than 0.05 it is not a normal
distribution we believe that the variable does not follow the normal distribution,
therefore use a nonparametric statistic.

Parametric Tests (p>0.05)

● Make assumptions about the parameters of the population distribution from which the sample
is drawn. If we draw a particular sample from a particular population, we assume that the data
is normally distributed
● Therefore, When data is normally distributed use parametric tests

Non-Parametric Test (p<0.05)

● The parameters of the population is not known and it is not normally distributed
● After doing the test of normality if the sample is not normally distributed we us a
nonparametric test
● If we don't know anything about the sample or what the normality looks like it is safe
to go for non parametric test

Measures of Central Tendency

In statistics, a central tendency is a central or typical value for a probability distribution. It

may also be called a center or location of the distribution. Measures of central tendency help
you find the middle, or the average, of a data set. The 3 most common measures of central
tendency are the mean, median and mode. The mode is the most frequent value. The median
is the middle number in an ordered data set.

What is Kurtosis?

Kurtosis is a statistical measure used to describe the degree to which scores cluster in the tails
or the peak of a frequency distribution. The peak is the tallest part of the distribution, and the
tails are the ends of the distribution.

What are the 3 Types of Kurtosis?

There are three types of kurtosis: mesokurtic, leptokurtic, and platykurtic.

Mesokurtic: Distributions that are moderate in breadth and curves with a medium peaked
height.
Leptokurtic: More values in the distribution tails and more values close to the mean (i.e.
sharply peaked with heavy tails)

Platykurtic: Fewer values in the tails and fewer values close to the mean (i.e. the curve has a
flat peak and has more dispersed scores with lighter tails).

Negative and Positive Value Kurtosis

Negative values of kurtosis indicate that a distribution is flat and has thin tails. Platykurtic
distributions have negative kurtosis values. Normally platykurtic.

Positive values of kurtosis indicate that a distribution is peaked and possess thick tails.
Leptokurtic distributions have positive kurtosis values. Normally leptopkurtic.

When kurtosis is equal to 0, the distribution is mesokurtic. This means the kurtosis is the
same as the normal distribution, it is mesokurtic (medium peak).

Null Hypothesis and Alternative Hypothesis

The null hypothesis states that there is no relationship between the two variables being
studied (one variable does not affect the other). It states the results are due to chance and are
not significant in terms of supporting the idea being investigated. Thus, the null hypothesis
assumes that whatever you are trying to prove did not happen.

The alternative hypothesis is the one you would believe if the null hypothesis is concluded to
be untrue. The alternative hypothesis states that the independent variable did affect the
dependent variable, and the results are significant in terms of supporting the theory being
investigated (i.e. not due to chance).

Measures of Dispersion

In the previous section we have discussed about measures of central tendency. By knowing
only the mean, median or mode, it is not possible to have a complete picture of a set of data.
Average does not tell us about how the score or measurements are arranged in relation to the
center.

It is possible that two sets of data with equal mean or median may differ in terms of their
variability.

Therefore, it is essential to know how far these observations are scattered from each other or
from the mean. Measures of these variations are known as the 'measures of dispersion'.

The most commonly used measures of dispersion are range, average deviation, quartile
deviation, variance and standard deviation.

Range: Range is one of the simplest measures of dispersion. It is designated by 'R'. The range
is defined as the difference between the largest score and the smallest score in the
distribution.

It is known as distance between the highest and the lowest scores in a distribution. It gives the
two extreme values of the variable but no information about the values in between the
extreme values. A large value of range indicates greater dispersion while a smaller value
indicates lesser dispersion among the scores.

Quartile Deviation : Quartile deviation is denoted as Q. It is also known as interquartile

range. It avoids the problems associated with range. Interquartile range includes only 50% of
the distribution.

Quartile deviation is the difference between the 75% and 25% scores of a distribution. 75th
percentile is the score which keeps 75% score below itself and 25th percentile is the score
which keeps 25% scores below itself.

Interquartile range is defined as the difference between the 25th and 75th percentile (also
called the first and third quartile). Hence the interquartile range describes the middle 50% of
observations.

Standard deviation is the most stable index of variability. In the computations of average
deviation, the signs of deviation of the observations from the mean were not considered.

In order to avoid this discrepancy, instead of the actual values of the deviations we consider
the squares of deviations, and the outcome is known as variance.

Further, the square root of this variance is known as standard deviation and designated as SD.
Thus, standard deviation is the square root of the mean of the squared deviations of the
individual observations from the mean.

How Do You Know If A "p-value" Is Statistically Significant?

When you perform a statistical test a p-value helps you determine the significance of your
results in relation to the null hypothesis. The level of statistical significance is often
expressed as a p-value between 0 and 1. The smaller the p-value, the stronger the evidence
that you should reject the null hypothesis. A p-value less than 0.05 (typically ≤ 0.05) is
statistically significant. It indicates strong evidence against the null hypothesis, as there is
less than a 5% probability the null is correct (and the results are random). Therefore, we
reject the null hypothesis, and accept the alternative hypothesis. However, this does not mean
that there is a 95% probability that the research hypothesis is true. The p-value is conditional
upon the null hypothesis being true is unrelated to the truth or falsity of the research
hypothesis. A p-value higher than 0.05 (> 0.05) is not statistically significant and indicates
strong evidence for the null hypothesis. This means we retain the null hypothesis and reject
the alternative hypothesis. You should note that you cannot accept the null hypothesis, we can
only reject the null or fail to reject it.

A statistically significant result cannot prove that a research hypothesis is correct (as this
implies 100% certainty). Instead, we may state our results “provide support for” or “give
evidence for” our research hypothesis (as there is still a slight probability that the results
occurred by chance and the null hypothesis was correct – e.g. less than 5%).

What is Skewness?
In probability theory and statistics, skewness is a measure of the asymmetry of the probability
distribution of a real-valued random variable about its mean. The skewness value can be
positive, zero, negative, or undefined.

It is the degree of distortion from the symmetrical bell curve or the normal distribution. It
measures the lack of symmetry in data distribution. It differentiates extreme values in one
versus the other tail. A symmetrical distribution will have a skewness of 0.

Types of Skewness

There are two types of Skewness: Positive and Negative.

Positive Skewness means when the tail on the right side of the distribution is longer or fatter.
The mean and median will be greater than the mode.

Negative Skewness is when the tail of the left side of the distribution is longer or fatter than
the tail on the right side. The mean and median will be less than the mode.

A symmetrical distribution will have a skewness of 0.

Skew Normality and Extremes

If the skewness is between -0.5 and 0.5, the data are fairly symmetrical.

If the skewness is between -1 and -0.5(negatively skewed) or between 0.5 and 1(positively
skewed), the data are moderately skewed.

If the skewness is less than -1(negatively skewed) or greater than 1(positively skewed), the
data are highly skewed.

Testing the Difference between 2 Means

Parametric Tests:

1. Independent Sample T-Test

a. Test is used to determine if two groups are significantly different from each
other
b. Used when the independent variable is a nominal or ordinal level of
measurement because it has categories and therefore it is called categorical
variable.
c. The dependent variable is in interval or ratio level of measurement or should
be continuous variables.
d. There is homogeneity of variance. Scores of the dependent variable in each of
the categories should have homogeneity in variance
e. The independent sample T test is used to check the difference of means in a
normal sample order which follows the rules of normality
 f. If the sample does not follow the rules of normality, that is, it is not a bell
shaped curve don't use the independent sample T test
g. Reject the null hypothesis if p < 0.05

Assumptions/ When? Where? What cases do we use this test?

i. Continuous Variable
➢ Must have a score of its own
➢ Mean value is continuous (interval or ratio) and is normally
distributed
ii. Should be normally distributed because it is a parametric test
iii. Should have a random sample
iv. Needs to have enough data a minimum of 5
v. Similar spread between group: the variance between the two groups,
that is, the standard deviation should be more or less similar

2. Paired Sample T-Test

a. Compare the means between two related groups on the same continuous
dependent variable
b. This is also called the dependent T test
c. The groups are related
d. Used to test within the group research hypothesis with pretest-posttest design
e. The independent variable is manipulated by the experimenter and examines its
effect on the dependent variable. Changes to the dependent variable is
assessed by comparing the pre test scores and post test scores.
f. Experimental condition is the independent variable and the positive and
negative effect is the dependent variable
g. Here p-value indicates the difference between pretest and posttest
Assumptions
i. Measured on a continuous scale (interval or ratio scale)
ii. The independent variable should consist of two categorical related
group or matched pairs
iii. There should be no significant outliers in difference between the two
related group
iv. Normally distributed
Interpretation
i. The p-value is less than the alpha level, p < 0.05
ii. We either accept or reject the null hypothesis accordingly if there is no
difference between the means
Non parametric Test:

1. Mann - Whitney U Test

a. Used to compare whether there is a difference in the dependent variable for
two independent groups
b. This is a nonparametric test where the median is compared.
c. If the p-value is less than or equal to 0.05 then the result is statistically
significant.
d. This table shows the difference.

2. Wilcoxon Rank Sum Test

a. Non parametric equivalent to the dependent T test
b. The test does not assume normality in the data
c. Used when the assumption has been violated and the use of dependent t-test is
inappropriate
d. When p > 0.05, we accept the null hypothesis depending on what the null
hypothesis states.
e. For example 0.071 is more than 0.05 therefore it is not statistically significant

Correlation

Q. What do we use to find the difference between 2 groups?

A. Generally, we use the t-test or the non-parametric versions of the t-test or anova test etc.

Correlation
It is a way of understanding the relation between two variables. (so, we basically run this to
see if there is a relationship between two variables)
- With this, we study only relationship - not the effect/impact
- There is no IV or DV in correlation.
- Once a relationship is established then there are multiple other statistical tools that we
used to see whether there is an effect (mediating effect or moderating effect) or an
impact between the variables.

(To understand) How do we test these?

By analysing if there is a relationship between:
- Children’s IQ & parents’ IQ
- Degree of social trust and no. of membership in voluntary associations
- Urban growth and air quality violations
- GRA funding and number of publications by Ph.D students
- No. of police patrols and no. of crimes
- Grade on exam and time on exam
- Intelligence and academic achievement
So, once the relationship is established there are multiple other statistical tools that we use to
see whether there is an impact, a mediating effect, a moderating effect, etc.

Correlation coefficient - ‘r’

It is the statistical index of the degree to which two variables are associated or related.
- With the help of this, we can determine whether 1 variable is related to the other
variable by seeing whether scores on the two variables covary --- whether they vary
together.
- When did run the test correlation we will get a statistical measure called Coefficient
correlation, this indicates the degree to which two variables are related, denoted by
the symbol r.

Direction of the relationship - This is the basic principle behind correlation

Variables can be +vely or -vely correlated.
● +ve correlation: value of 1 variable ↑, value of other variable ↑.
● -ve correlation: value of 1 variable ↑, value of other variable ↓.

Variables vary at the same level. We can determine whether one variable is related to another
by seen whether scores on the two variables covary, that is, whether the vary together
● Eg: variable A ↑, variable B ↓ - in equal measures

Strength of the relationship (correlation)

1. Is indicated by its numerical value

- Anything closer to 0 is a weak correlation
- 0 means no correlation
2. Is not denoted by the sign, therefore, the sign must be ignored (i.e. the direction of the
relationship). Strength of the correlation is the same despite the sign.
3. Expresses the strength of the linear relationship between the variables. A correlation
of 0 indicates the value is not linearly related.

Interpreting the correlation output

- ‘r’ ranges from -1 to +1 (strongest -ve correlation is -1)

- Signs indicate the directions
- The numerical value indicates the strength
- Perfect correlation: -1 or +1
- No correlation: 0
- A correlation of 0 indicates that the values are not linearly related.
- Range: +1.0 to -1.0
 - Above +0.5 indicates high correlation
- Closer to 0 indicates weak correlation

Pearson’s Correlation Coefficient - parametric

- go with this when the data is normal
(This has a non-parametric equivalent.)

The Pearson product moment correlation coefficient (Pearson’s correlation) is a measure of

the strength and direction of association that exists between 2 variables measured on at least
one an interval scale.
Eg. - This can be used to understand whether there is an association between exam
performance and time spent revising.
- This can also be used to understand whether there is an association between depression
and length of unemployment.

Assumption of the test

1. Variables should be measured on an interval or ratio scale. It should be on a

continuous scale
2. There should be no significant outliers, however it can have one or two outliers but it
should not have significant outliers.
3. The data should follow normal distribution
4. A linear relationship between variable needs to be present

Extra IMP points

● Find the difference between two groups conduct a T test
● To establish relationship between two variables conduct a correlation test
● There is no independent variable and dependent variable in correlation

(What are we looking at? Relationship)

Explanation:
● Graph 1 - Linear:
The graph will follow a straight line on plotting the coordinates: This means that there
is a similar variance.
Eg.: If 1 variable ↑ by 1 unit, the other would possibly ↑ or ↓ by 1 unit.

*normal distribution doesn’t mean that you won’t have outliers at all (there are outliers there
also)

EXAMPLE of a table showing Pearson’s Product Moment Correlation:

Interpretation from the table:

- Pearson’s correlation: 0.706 denotes strong positive correlation. (why? Because it’s
above 0.5 and has a positive sign)
- It is significant as well. (Since the significance value is <0.05)
Significant outliers (Brief Information):
- Significant outliers denote any data point which is extremely high or extremely low
from that of the mean.
- Easiest way to figure out whether there are significant outliers or not is to run the test
of normality.

Spearman Rank Order Correlation

- Non parametric version of Pearson’s Correlation Coefficient (data is not normally
distributed)

The Spearman rank order correlation coefficient (Spearman’s correlation) is a non-parametric

measure of the strength and direction of association that exists between 2 variables measured
on at least an ordinal scale.
- Is denoted by the symbol rs (or the Greek letter ρ, pronounced rho) (the s is a
subscript)
- The test is used for either ordinal variables or for continuous data that has failed the
assumptions necessary for conducting the Pearson’s product moment correlation.

Interpretation from the table: SAME AS PEARSON’S

General interpretation:
- ≥ 0.5 correlation
- 0.3 to 0.5 correlation
- ≤ 0.3 weak correlation

Scatter Plots
Correlations between quantitative variables are often presented using scatterplots. Each point
in the scatter plot represents one person's score on both variables.

Taking all the points into account, one can see that people under more stress tend to have
more physical symptoms.

This is a good example of a positive relationship, in which higher scores on one variable tend
to be associated with higher scores on the other.

A negative relationship is one in which higher scores on one variable tend to be associated
with lower scores on the other.
KENDALL'S TAU (T) -

Kendall's tau is another useful measure of correlation. It is as an alternative to Spearman's rho

(rs ). This correlation procedure was developed by Kendall (1938). Kendall's tau is based on
an analysis of two sets of ranks, X and Y.

The range of tau is from - 1.00 to + 1.00. The interpretation of tau is based on the sign and
the value of coefficient.

The tau value closer to ‡1.00 indicates stronger relationship.

Positive value of tau indicates positive relationship and vice versa.

BISERIAL CORRELATION (RPB)

Some variables are dichotomous.

The dichotomous variable is the one that can be divided into two sharply distinguished or
mutually exclusive categories.

Some examples are, male-female, rural-urban, Indian-American, diagnosed with illness and
not diagnosed with illness, Experimental group and Control Group, etc.

These are the truly dichotomous variables for which no underlying continuous distribution
can be assumed.

Point Biserial Correlation (rpb) is Pearson's Product moment correlation between one truly
dichotomous variable and other

continuous variable. Algebraically, the rpb = r. So we can similarly calculate rpb.

POINT BISERIAL CORRELATION (RPB)

Some variables are dichotomous.

The dichotomous variable is the one that can be divided into two sharply distinguished or
mutually exclusive categories.

Some examples are, male-female, rural-urban, Indian-American, diagnosed with illness and
not diagnosed with illness, Experimental group and Control Group, etc.

These are the truly dichotomous variables for which no underlying continuous distribution
can be assumed.

Point Biserial Correlation (rpb) is Pearson's Product moment correlation between one truly
dichotomous variable and other

continuous variable. Algebraically, the rpb = r. So we can similarly calculate rpb.

PHI COEFFICIENT

When both the variables are dichotomous, then the Pearson's correlation calculated is called
as Phi Coefficient (Ф).

For example, let us say that you have to compute correlation between gender and ownership
of the property.

The gender takes two levels, male and female.

The ownership of property can be measured as either the person owns a property and the
person do not own property.

Now you have both the variables measured as dichotomous variables.

Now if you compute the Pearson's correlation between these two variables is called as Phi
Coefficient (ф ). Both the variables take value of either of 0 or 1.

TETRACHORIC CORRELATION (RTET)-

Tetrachoric correlation is a correlation between two dichotomous variables that have

underlying continuous distribution.

If the two variables are measured in a more refined way, then the continuous distribution will
result. For example, attitude to females and attitude towards liberalisation are two variables to
be correlated.

Now, we simply measure them as having positive or negative attitude. So we have 0

(negative attitude) and 1 (positive attitude) scores available on both the variables.

Then the correlation between these two variables can be computed using Tetrachoric
correlation (rtet).
 Null & Alternative Hypothesis
Hypothesis
A hypothesis is an appropriate explanation that relates to the set of facts that can be tested by
certain further investigations.

Two types:
1. Null Hypothesis
2. Alternate hypothesis

A research generally starts with a problem. Hypothesis provides the researcher(s) with some
specific restatements and clarifications of the research problem. It is an assumption based on
the existing literature or facts

Important points regarding null hypothesis and alternate hypothesis :

- provide direction to the research

- The criteria of the research problem in the form of null hypothesis and alternate
hypothesis should be expressed as a relationship between two or more variables.
(whether the two variables are related or not)
- The criteria is that the statements should be the ones that express the relationship
between two or more measurable variables. (extra: in qualitative research, we don't
make use of such hypotheses)
- Both the null hypothesis and the alternate hypothesis should carry clear implications
for testing and stating relations. (so, they should clearly indicate what kind of testing
measure we would be using)
- The major differences between the null hypothesis and alternate hypothesis and the
research problems are that the research problems are simple questions that can not be
tested, whereas these hypotheses can.
- Framing of the hypothesis has to be very clear and concise. The null hypothesis and
alternate hypothesis are required to be framed properly before the data collection and
interpretation phase in the research.
- Well-framed hypotheses indicate that the researcher has adequate knowledge in the
particular area and is thus able to take the investigation further because they can use a
much more systematic system.
- It gives direction to the researcher on his/her collection and interpretation of data.
- The null hypothesis and alternate hypothesis are useful only if they state the expected
relationship between the variables or if they are consistent with the existing body of
knowledge.
- The purpose is to provide the researcher or an investigator with a relational statement
that is directly tested in a research study.
- The purpose and importance of null hypothesis and alternate hypothesis are that they
provide an appropriate description of the phenomena.
 - The purpose is to provide the framework for reporting the inferences of the study.
- The purpose is to behave as a working instrument of the theory.
- The purpose is to prove whether or not the test is supported which is separated from
the investigator’s own values and decisions.

Null hypothesis

- Generally denoted as H0
- States the exact opposite of what an investigator or an experimenter predicts or
expects.
- It basically defines the statement which states that there is no exact or actual
relationship between the variables. - we assume that there will be no outcomes
- Here, no bias is involved. (hence it it typically preferred)

Alternate hypothesis

- Generally denoted as H1
- Makes a statement that suggests or advises a potential result or an outcome that an
investigator or the researcher may expect. (typically used when there is a strong
relationship between the variables)
- Has been categorised into two categories:
1. Directional alternative hypothesis
2. Non-directional alternative hypothesis
 Answer: 1st option

Extra
When a curve is presented:
1. hightest value is towards the right side - tail 2
2. lowest value is towards the left side - tail 1
3. moderate/average is the middle

H0: fr = fa (friends and family have equal understanding/no difference in understanding me)
- Non-directional hypothesis because it assumes that there is no difference between
the direction of tail 1 as well as the direction of tail 2 - there is no significant
difference between the two curves, specifically the tails.
- Both curves are identical or almost similar
- Also called two tailed hypothesis because we have to compare both the tails along
with the middle portion of the curve to draw a conclusion

HA: fr ≠ fa
- An alternative of the null hypothesis
- Chosen when null hypothesis is rejected, therefore it is an alternative hypothesis
- Also called two tailed hypothesis because we are comparing tail 1 and 2 in both
cases
- It is bi-directional as the hypothesis indicates a direction towards tail 1 and tail 2

HA: fr > fa (friends understand me better than family)

- Compare only one tail for both curves (each curve represents data collected by family
and friends respectively)
- Alternative hypothesis aka directional hypothesis - clearly points to a direction
- Since we are stating better/greater than, only look at moderate to high
 A statistically significant difference vs a practically important
difference

Statistical significance

- It refers to the unlikelihood that mean differences observed in the sample have
occurred due to sampling error. - It is not ideally possible to have our sample
represent an entire population and subject them to a test, hence even in a large sample
size, there might still be the possibility of an error.
- Given a large enough sample, one might still find statistical significance.
- It rules out the possibility of any sort of sampling error.
- With a larger sample size, there is high likelihood of statistical significance. (because
statistical significance is always calculated based on the sample size by the software
---- because the larger the sample is the difference for each data point is being shown,
therefore, when every data point has a minute change also, it will be regarded as
statistically significant)
*Whatever results we get by testing a particular sample, we tend to generalise those to the
entire general population.

What does ‘statistically significant’ mean?

- By using the term significant, we aren't automatically implying that the data is
important.
- It is simply a term that is used to assess whether the evidence against the H0 has
reached the standard set by α only. (we usually set the confidence interval at 95%
(sometimes 99%) level, i.e. P ≤ 0.05, so, α = 0.05)

When we say that it (the result/data) is statistically significant, what we mean is that we have
a preset α level, i.e. 0.05. So what we imply is that, according to this level, by testing the H0,
we found that this particular difference is statistically significant.
- For example: Significance level at 0.05 is often expressed by the statement: “The
results were significant at (P< 0.05)”, where P stands for the P-value. If P is less than
0.05, the result is statistically significant.
- The P-value is helpful in providing basic information rather than a statement of
significance, because we can then assess significance at any level we choose.
- So, it is providing some information about the sample at a level that we choose as
researchers. It is actually not conveying the strength or the meaningfulness of the
relationship or the difference. It’s a purely statistical measure.
Practical significance

- It looks at whether the difference is large enough to be of value in a practical sense.

The aim of any research is to provide contribution to any existing research/literature
or to existing scientific practises.
- Practical significance refers to the importance or usefulness of the result in some
real-world context.
- Many sex-differences are statistically signifcant — and many may even be interesting
for purely scientific reasons — but they are not practically significant.

- It is not the same as statistical significance. If a result is statistically significant, it

does not imply that the result is practically significant
- Due to this problem, many editors and publishers are requiring authors to include
some information surrounding the practical significance of their findings. (because
only then is there a point in disseminating the findings)
- APA guidelines require reporting of effect sizes wherever possible.

Statistical significance alone can be misleading because it’s influenced by sample size. (it
alone can not predict any kind of generalizability)
- Increasing the sample size always makes it more likely to find a statistically
significant effect, no matter how small the effect truly is in the real world. But does
that actually mean that the relationship is strong & meaningful?
- Measuring effect sizes is a way in which we can assess whether the values or the
results are actually practically significant.
- In contrast, effect sizes are independent of the sample size. Only the data is used to
calculate the effect sizes. So, this actually measures the strength or the
meaningfulness of the relationship.

Effect size on statistical significance

- When a difference is statistically significant, it does not necessarily mean that it is big,
important or helpful. It simply means that you can be confident that there is a
difference.
- Effect size (usually denoted by Cohen’s d) is a measure of the strength of the
relationship between 2 variables.
- In research settings, it is not only helpful to know whether results have a statistically
significant effect, but also the magnitude of any observed effects.
- To know if an observed difference is not only statistically significant but also
important or meaningful, you will need to calculate its effect size.
- When H0 (“no effect” or “no difference”) is rejected at the usual levels (α = 0.05 or α
= 0.01), there is a good evidence that an effect is present. However, that effect may be
extremely minor or trivial.
 - When a larger sample size is accessible, even tiny deviations from the H0 will be
significant.
- Today, whenever there’s an intervention/treatment provided to the experimental
group, it is important to mention the effect size and to report any sort of effect that the
experimenter observed in the participants or in the sample. For example, if the
participants in the experimental group report that their stress levels were reduced due
to the intervention, this acts as evidence that the intervention provided to the
experimental group has had an effect on the participants and this further contributes to
the research being conducted.
Extra:
- 0.1 to 0.3 is low effect size
- 0.3 to 0.5 is moderate effect size
- and ≥ 0.5 is high effect size
- H0 = null hypothesis
- HA = alternative hypothesis

--------------------------------------------------------------

Significant study results vary based on context. Can significant study results ever be
translated into recommendations for the general population?
We know so far that:
1. Tests of statistical significance rarely tell us about the importance of a research result.
2. Effect size tells us about magnitude of difference, which is important, but it is difficult
for practice-oriented practitioners to comprehend.

What can we do, otherwise, to make some sense out of the statistics? We may not do effect
sizes all the time. Does that mean that if you get statistically significant results, you discard
them or you disbelieve them? NO.

Solutions? (basically how to make sense of the statistically significant data in such a
situation?)
There are several solutions to help translate statistically significant data into results that may
be practically applied to real life situations:
1. Comparison of a sample to a meaningful reference group (standardised proper norms)
2. Confidence limits (helps establish generalisability of your result(s))
3. Know when to detect faulty data (one way of assessing this is to see whether the
sample is representative of the population -- think -- “am I trying to communicate
something which can be easily generalized or something which is specific to my
sample or study?”)

Levels of significance vs. p-values

Choosing a level of significance

- Certain standard levels of significance are often used such as 10%, 5%, and 1%.
- The 5% level (α= 0.05) is particularly common. (but in intervention it is always better
to supplement that with effect sizes)
- Significance at the 5% level is still a widely accepted criterion for meaningful
evidence in research work.

- Basically, p-value can be 0.04 or 0.051. If it is 0.04, you will mention it as statistically
significant, while the latter as not significant. So, actually the difference is very
minimal but just based on that particular value that you have set, you interpret the data
accordingly. So, it’s not a very reliable measure in terms of understanding the
effectiveness or the practical significance of your data.
 - If the results are not statistically significant, it does not mean that they are
meaningless. We just have to further evaluate to land on a conclusion (to confirm the
finding(s)).
- So any scientific enquiry has its own meaning, in its own form. The only way is, the
more you structure it, the better it becomes. The more generalizable it is, the more
meaningful it is.

Types of Experimental Designs

Broadly there are two types:

1. Informal experimental designs

● Designs that normally use a less sophisticated form of analysis based on

differences in magnitudes.
● Have lesser scientific rigor (do not follow all the processes that need to be
done in an experiment)

Major Informal Experimental Designs include:

1. Before-and-after without control design.

2. After-only with control design.
3. Before-and-after with control design.

2. Formal experimental designs

● Offer relatively more control and use precise statistical procedures for
analysis.
● Variables are more controlled
● These allow the interpretation of results with much more confidence.
● Here, we control most of the variables which we think might affect the
cause-effect relationship between the IV and DV.

*Conducting experiments includes controlling many extraneous variables or confounding

variables so as to ensure that we clearly oversee and interpret the cause & effect relationship
between two variables or more. So, if other variables are influencing the experiment, then as
an experimenter we may have to take such a situation into account. The use of formal
experimental designs, thus, helps ensure that the variables are more controlled.

Major Formal Experimental Designs include:

1. Completely randomized design (C.R. Design).

2. Randomized block design (R.B. Design).
3. Latin square design (L.S. Design).
4. Factorial designs.
 *Good experimental studies usually use Formal Experimental Designs. However, making
a choice between the two types of designs totally depends on the research problem.

Experimental Research Design

1. Weak Experimental Designs

A. One-Group Posttest-Only Design

i. Only one post test is being conducted
ii. Treatment condition is also IV
iii. Cannot build any cause - effect relationship based on this. Only
qualitative data can be produced.
iv. No control group (extraneous factors are controlled, not given any
intervention)

B. One-Group Pretest - Posttest Design

i. There will be pretest and post test, but there’s only one group -
experimental group/treatment group (treatment grp gets the
intervention)
ii. No control group

C. Posttest - Only with Nonequivalent Groups Design

i. Generally, there are at least two groups (treatment group and a control
group), one of which does not receive a treatment or intervention
ii. Data are collected on the outcome measure after the treatment or
intervention
 iii. Control group and experimental group have different numbers of
participants. (eg: 10-20, 15-17, etc.)

2. Strong Experimental Designs

A. Between Participants - Posttest Only Control Group Design

i. To assess a particular variable between two groups (experimental &
control)
ii. Both groups have equal number of participants, therefore (because of
equal number of participants and the presence of a control group), it is
a strong experimental design.
iii. May not do a pretest but is a strong experimental design
iv. Can have levels
v. Placebos are also given; placebo groups also exists
vi. Conceptually simpler and require less testing time per participant
vii. Useful in avoiding carryover effects without the need for
counterbalancing
viii. Preferred because it will not tax/drain/ask much from the participants.

B. Within Participants - Within Participants Posttest Only Design

i. Only one group
ii. All participants go through both conditions; experimental and control
group conditions
iii. Controlling extraneous participant variables, which generally reduces
noises in the data
iv. Makes it easier to detect a relationship between the IV and DV
v. Usually used to understand an effect of a particular variable on both
conditions

Difference between: Between Participant and Within Participants?

In a between-subjects design:
- Also known as an independent measures design or classic ANOVA design
- Individuals receive only one of the possible levels of an experimental treatment.
- In one might use matched pairs within the between-subjects design to make sure that
each treatment group contains the same variety of test subjects in the same
proportions.
In a within-subjects design:
- Also known as a repeated measures design
 - Every individual receives each of the experimental treatments consecutively, and their
responses to each treatment are measured.
- Within-subjects or repeated measures refer to an experimental design where an effect
emerges over time, and individual responses are measured over time in order to
measure this effect as it emerges.

C. Mixed - Pretest - Posttest Control Group Design

i. Two groups
ii. Both go through the pretest and posttest

D. Factorial
i. Can have between participants, within participants and mixed model

Effects :

Qualitative Research :
Qualitative research can be defined as a type of scientific research that tries to bridge the gap
of incomplete information, systematically collects evidence, produces findings and thereby
seeks answer to a problem or question.
It is widely used in collecting and understanding specific information about the behaviour,
opinion, values and other social aspects of a particular community, culture or population.
An example of a qualitative research can be studying the concepts of spiritual development
amongst college students.

There are certain approaches to qualitative research.

1. Case study: Helps in providing an in depth knowledge of the nature, process or phenomena
of a specific case under study.
Multiple methods of data collection are often used in case study research (example,
interviews, observation, documents, and questionnaires).
2. Ethnography: This approach mainly focuses on a particular community. It is more of a kind
of closefield observation and basically tries to study a socio cultural phenomena.
Ethnography can be used for comparative analysis of cultural groups (e.g. eating habits of
North Indians and South Indians), also known 'Ethnology'. Further it can also be used to
analyse the cultural past of group of people (e.g.
Harappan civilisation), also known as 'Ethnohistory'.

3. Historical method: This method helps in understanding and analysing causal relationships.
Data related to the occurrence of an event is collected and evaluated in order to understand
the reasons behind occurrence of such events.
It helps in testing hypothesis concerning cause, effects and trends of events that may help to
explain present events and anticipate future events as well.
4. Grounded theory: This approach involves an active participation of the researcher in the
activities of the group, culture or the community under study.
The data regarding the required information is collected with the help of observation._
It is generally used in generating or developing theories. This means that the ground theorists
can not only work upon generation of new theories, they can test or elaborate previously
grounded theories.

5. Phenomenology: In this method, behavioural phenomena are explained with the help of
conscious experience of events, without using any theory, calculations or assumptions from
other disciplines.
The concept can be best understood with the help of one of the studies that was done in which
patients were asked to describe about caring and uncaring nurses in hospitals Creswell, 1998.
The patients explained those nurses to be caring who show their existential presence and not
mere their physical presence. The existential presence of caring nurses referred to the positive
response showed by them to the patient's request.
The relaxation, comfort and security that the client expresses both physically and mentally
are an immediate and direct result of the client's stated and unstated needs being heard and
responded to by the nurse.

6. FOCUS GROUPS > 8-10 people

In this method of qualitative research, certain questions or issues are discussed by a group of
participants.
Focus groups can be structured or unstructured and the group could be heterogeneous or
homogeneous in nature.
It may involve experts (referred to as Delphi group) who discuss certain important issues (for
instance, health educators could discuss about certain health related issues of a certain
community) or could be individuals from certain community/ area discussing about certain
issue (like women empowerment).
The size of the group will depend on the objective(s) of the research (though an ideal size
would be twelve and for Delphi groups it could be around four participants) and the
participants may or may not know each other.
The main advantages of this method is that a lot of information is gathered.
Though, there are disadvantages as well, as there could issues related to encouraging the
participants to contribute equally and actively.
There is also the problem of social desirability on part of the participants that can impact the
outcome of the study.

7. ACTION RESEARCH
Action research is "learning by doing" - a group of people identify a problem, do something
to resolve it, see how successful their efforts were, and if not satisfied, try again. While this is
the essence of the approach, there are other key attributes of action research that differentiate
it from common problem-solving activities that we all engage in every day.
What separates this type of research from general professional practices, consulting, or daily
problem-solving is the emphasis on scientific study, which is to say the researcher studies the
problem systematically and ensures the intervention is informed by theoretical
considerations. Much of the researcher's time is spent on refining the methodological tools to
suit the exigencies of the situation, and on collecting, analyzing, and presenting data on an
ongoing, cyclical basis.
Action research became popular in the 1940s. Kurt Lewin (1946) was influential in spreading
action research. He came interested in helping social workers improve their practice.
Participatory action research (PAR) is a special kind of action research in which there is
collaboration between the study participants and the researcher in all steps of the study.

8. NARRATIVE RESEARCH
Narrative research aims to explore and conceptualize human experience as it is represented in
textual form. Aiming for an in-depth exploration of the meanings people assign to their
experiences, narrative researchers work with small samples of participants to obtain rich and
free-ranging discourse.
The emphasis is on storied experience. Generally, this takes the form of interviewing people
around the topic of interest, but it might also involve the analysis of written documents.
Narrative psychology is concerned with the structure, content, and function of the stories that
we tell each other and ourselves in social interaction. It accepts that we live in a storied world
and that we interpret the actions of others and ourselves through the stories we exchange.
Through narratives we not only shape the world and ourselves but they are shaped for us
through narrative.
People who have undergone some sort of trauma provide instances of how they often try to
make sense of the events that they are going through by creating stories or narratives.
A narrative is essentially a written or spoken account of connected events with an underlying
time dimension.
Designing an Experiment

An experiment is a type of research method which is used to manipulate one or more

independent variables and measure their effect on one or more dependent variables. An
experimental design involves creating a unique set of procedures to test the hypothesis
proposed by the researcher.

Step 1: Asking a research question

Framing an appropriate research question involves

● logic behind it
● how it needs to be measured
● Concrete
● Structured
● Specific
● And which can be studied
● shouldn’t be ambiguous

The primary purposes of formulating the problem in question form are to ensure that the
researcher has

● A good grasp of the variables to be investigated

● To aid the experimenter in designing and carrying out the experiment
● Specificity of the research question is an important consideration

To translate the research question into an experimental hypothesis, we need to define the
main variables and make predictions about how they are related. This involves extensive
research and identification of the independent variables and dependent variables.

Step 2: Operationalizing the variables

1. What the conditions they are being exposed too

2. The level of the independent variable/s are manipulated, or systematically varied.
➢ The different levels of the independent variable are called experimental conditions.
➢ There can be more than one experimental condition.
➢ Ex : Number of hours of sleep; Type of stimulus (visual, echoic, tactile) ; Amount of
exposure to aggressive models.

3. The variability in extraneous variables other than the independent and dependent
variable are controlled, or minimized.
 ➢ We need to be aware of the presence of the extraneous variables/factors that will
affect the study and the variables being studied in the research
➢ In an experiment, an extraneous variable is any variable that we’re not
investigating that can potentially affect the outcomes of our research study.
■ If left uncontrolled, extraneous variables can lead to inaccurate conclusions about
the relationship between independent and dependent variables.
■ Extraneous variables can threaten the internal validity of your study by providing
alternative explanations for your results.
■ In an experiment, you manipulate an independent variable to study its effects on
a dependent variable.
➢ In research that investigates a potential cause-and-effect relationship, a
confounding variable is an unmeasured third variable that influences both the
supposed cause and the supposed effect.
■ Confounding variables (a.k.a. confounders or confounding factors) are a type of
extraneous variable that are related to a study’s independent and dependent
variables.
■ A variable must meet two conditions to be a confounder:
● It must be correlated with the independent variable. This may be a causal
relationship, but it does not have to be.
● It must be causally related to the dependent variable.

4. Manipulation and control have different meanings in experiments.

➢ Researchers manipulate the independent variable by systematically changing its
levels and control other extraneous variables by holding them constant.
■ The extraneous factor setting needs to be monitored and held constant so that it is
not going to affect the cause effect relationship.

5. All the variables need to be operationalized

➢ Operationalization is the conversion from research question to experiment design.
➢ It means that the IVs and DVs are SMART – Specific, Measurable, Achievable,
Realistic and Timebound
■ Can be administered and measured in a given duration of time
➢ Operationalization reduces subjectivity and increases the reliability of the study.

Step 3: Framing the Hypothesis

A hypothesis is a statement that can be tested by scientific research.

● To test a relationship between two or more thing

● Write hypotheses before you start your experiment or data collection.
A hypothesis states your predictions about what your research will find. It is a tentative
answer to your research question that has not yet been tested.

One should ensure that the hypothesis written is a specific, testable hypothesis that addresses
the research question.

A hypothesis is not just a guess

● It should be based on existing theories and knowledge

● It also has to be testable, which means we must be able to support or refute it through
scientific research methods like experiments, observations and statistical analysis of
data

Hypotheses propose a relationship between two or more variables.

● An independent variable is something the researcher changes or controls.

● A dependent variable is something the researcher observes and measures.
● For example: Hypothesis - Daily apple consumption leads to fewer doctor’s visits.
○ The independent variable is apple consumption — the assumed cause.
○ The dependent variable is the frequency of doctor’s visits — the assumed
effect.

Developing a Hypothesis

1. Writing a hypothesis begins with a research question that we want to answer. The
question should be focused, specific, and researchable within the constraints of our
project.
2. Construct a conceptual framework to identify which variables you will study and what
you think the relationships are between them.

Step 4: Deciding the Design

Researchers design the experiment to answer their research question.

Design provides the framework of the study.

● Once established, this framework must be explained clearly and implemented.

● The researcher has to determine the kinds of participants to be used, from where they
can be obtained, how many should be asked to participate and what instructions and
tasks will be given.

Choice of design is based on the relative merits for a specific research question

● Design depends on the research question

Randomized or strong designs are the best experimental designs available.

● Randomization

An experiment can be completely randomized or randomized within blocks (aka

strata):

- In a completely randomized design, every subject is assigned to a treatment group at

random.
- In a randomized block design (aka stratified random design), subjects are first
grouped according to a characteristic they share, and then randomly assigned to
treatments within those groups.

Considerations for Step 4

1. Should I use a control group ? - Usually, in any randomized control trial or

experiment, it is ideal to use a control grp
2. Should I use multiple treatment comparison groups (comparing more than one active
treatment)
3. Should I use a pretest?
4. Should I use just one or multiple pretests (to get a stable baseline)? One pretest gets
the score, second is done to confirm it
5. Should I use just one or multiple posttests (to get a stable treatment effect or identify
6. delayed outcomes)? Yes. once a post test is done, they conduct a follow up test that is
after the post test, just to ensure that effect is still there/is still a factor
7. Should I use a within-subjects or a between-subjects independent variable, or should I
use both?
8. Should I include multiple theoretically interesting independent variables in the design
(as in factorial designs)?
9. Should I include more than one dependent variable (to see how the treatment affects
several different outcomes)?

Step 5: Asking will this experimental study be valid?

Researchers have focused on four validities to help assess whether an experiment is sound

1. Internal validity : Does changing the IV in two similar experimental conditions cause
a direct change in DV in each condition?
a. Only if the IV results in significant changes in the DV can we establish a
cause - effect relationship.
b. The way an experiment is conducted supports the conclusion that the
independent variable caused observed differences in the dependent variable. -
This proves that the experiment and it’s findings have a strong relevance.
 2. External validity : Do the situation and behaviours measured simulate real life
settings?
a. This is applicable in laboratory settings; will the results of a laboratory test be
relevant in a non-laboratory, real life setting - how do we generalise it to an
actual setting.
b. The way a study is conducted supports generalizing the results to people and
situations beyond those actually studied. - what were the variables controlled.
c. If it is a complete laboratory experiment that is artificial to real life, the
findings will not stay true in a real life setting.

3. Construct validity : Do the variables adequately measure the construct being studied?
a. This refers to the quality of the experiment’s manipulations.
b. Whether what we have measured is actually what we intend to measure
and whether the variables have been operationalised well.
c. Ex: Is the number of words read in one minute an accurate measure of reading
ability?

4. Statistical validity : Does the data demonstrate whether the difference or relationship
that was predicted was found?
a. To design a statistically valid experiment, thinking about the statistical tests at
the beginning of the design will help ensure the results can be believed
b. Have used the right test/method of analysis to help establish the cause and
effect relationship.

All the types of validity are ideally not possible in a single study.

Step 6:Determining the active role of the experimenter in the manipulation

● Manipulation of an independent variable must involve the active intervention of the

researcher.
● Comparing groups of people who differ on the independent variable before the study
begins is not the same as manipulating that variable.
● Thus the active manipulation of the independent variable is crucial for eliminating the
third-variable problem (that extraneous variables and not the IV could have caused
changes in the DV) - active manipulation occurs by the active role of the
experimenter, the experimenter causes it.
 Institutional Review board
● Prepare a research protocol that details all aspects of the research, including the type
of participants you propose to use and the procedures that will be employed in
conducting the study
● The primary concern of the IRB is the welfare of the human participants.
● Informed consent for participation in the study and that the procedures will not harm
the participants.
● Weighing risks vs benefits of the experimental study
● Cover all ethical aspects like privacy and confidentiality

Recruiting participants
● Every experimental study consists of participants drawn from a subject pool : An
established group of people who have agreed to be contacted about participating in
research studies.
● Researchers selecting humans as their research participants must decide on the
inclusion and exclusion criteria for their participants.
● What type of participants do you need? Ex: Users of public transport , A general
undergraduate population , persons undergoing therapy for mental health issues etc.
● What resources do I have to recruit participants? Are participants recruited based on
convenience and availability?
● The point of having a well-defined selection rule is to avoid bias in the selection of
participants.
● Alternatively, if your strategy is to post a research study on the Internet and have
participants log on to the Web site and complete the study, you could post the study on
one of several Web sites that specialize in advertising research opportunities.
● One of these sites is hosted by the Social Psychology Network, http://
www.socialpsychology.org/addstudy.htm, and another is hosted by the American
Psychological Society, http://psych.hanover.edu/research/exponnet.html.
● After identifying the target participant population, the researcher must select
individual participants from that group through random sampling.
● The number of participants needed to test the hypothesis adequately must be decided.
This decision is based on the
○ design of the study,
○ variability of the data
○ the type of statistical procedure to be used
● As the number of participants within a study increases, the ability of our statistical
tests to detect an effect of the independent variable increases; that is, the power of the
statistical test increases.

Sampling
 ● In research terms a sample is a group of people, objects, or items that are taken from a
larger population for measurement.
● The sample should be representative of the population to ensure that we can
generalise the findings from the research sample to the population as a whole.
● Probability sampling involves random selection, allowing you to make strong
statistical inferences about the whole group.
● Non-probability sampling involves non-random selection based on convenience or
other criteria, allowing you to easily collect data.

Probability Samping

In psychology research, sampling methods are crucial for gathering data that accurately
represents a larger population. Probability and non-probability sampling are two primary
approaches used to select participants for a study, each with distinct advantages and
limitations.

Probability Sampling:

Definition: Probability sampling involves selecting participants randomly from a larger

population, ensuring each member of the population has an equal chance of being chosen.
Common methods include simple random sampling, stratified sampling, cluster sampling,
and systematic sampling.

Advantages:

Provides a higher level of representativeness, as each member has an equal opportunity to be

included.

Results are more likely to be generalizable to the larger population, increasing external
validity.

Allows for statistical inferences and the calculation of margin of error.

Limitations:

May be more resource-intensive and time-consuming, especially for large populations.

Not always feasible due to logistical constraints or when a sampling frame (list of the entire
population) is unavailable.
Non-Probability Sampling

Non-Probability Sampling:

Definition: Non-probability sampling involves selecting participants through non-random

methods. This can include convenience sampling, purposive sampling, snowball sampling, or
quota sampling.

Advantages:

Convenient and cost-effective, especially when access to the entire population is difficult.

Useful for exploratory research or when specific types of participants are needed.

Can be quicker to implement, making it suitable for small-scale studies.

Limitations:

May result in a biased sample that doesn't represent the entire population accurately.

Findings might lack generalizability and external validity.

Difficult to measure the extent of sampling error or make statistical inferences.

In psychology, the choice between probability and non-probability sampling often depends on
the research goals, available resources, and the nature of the study. For instance, if the aim is
to generalize findings to a larger population, probability sampling is preferred. However, in
cases where specific types of participants are needed or access to the entire population is
challenging, non-probability sampling might be more practical
 Apparatus and Instruments
● Researchers must identify how the independent variable conditions will be presented
(computer based software) and how the dependent variable will be measured
(subject’s verbal response to record speed, accuracy, recall etc.)
● In some experimental studies, the presentation and manipulation of the independent
variable requires the active participation of the investigator, and the measurement of
the dependent variable involves the administration of a variety of psychological
assessment instruments.

Standardizing the Procedure : Establishing controls

● Extraneous variables may be unknowingly introduced during the procedure how
participants are to be greeted, type of nonverbal behavior (looking at the participants,
smiling, using a particular tone of voice in reading instructions, etc.) as well as the
verbal behavior in which the experimenter is to engage.
● This can add noise to the data and make the effect of the independent variable more
difficult to detect. If they vary across conditions, they become confounding variables
and provide alternative explanations for the results.
 A checklist to minimize variation in procedure
● Create a written protocol that specifies everything that the experimenters are to do and
say from the time they greet participants to the time they dismiss them.
● Create standard instructions that participants read themselves or that are read to them
word for word by the experimenter.
● Automate the rest of the procedure as much as possible by using software packages
for this purpose or even simple computer slide shows.
● Anticipate participants’ questions and either raise and answer them in the instructions
or develop standard answers for them.
● Train multiple experimenters on the protocol together and have them practice on each
other.
● Be sure that each experimenter tests participants in all conditions.

Instructions
● “What should be included in the instructions?” and “How should they be presented?”
● Instructions must include
○ a clear description of the research purpose, or disguised purpose
● Instructions should be clear, simple, unambiguous, and specific, but at the same time
they should not be too complex
● Include “warm-up” trials as part of your instructions. These are pretest trials that are
similar to those the participant would complete in the actual study. They are included
to ensure that the research participant understands the instructions and the way they
are to respond.
● Instructions requesting that the research participant “pay attention,” “relax,” or
“ignore distractions” are probably ineffective because research participants are
constrained by other factors that limit their ability to adhere to the commands.
● Instructions sometimes request that the participants perform several operations at the
same time. If this is not possible, then they will choose one of the possible operations
to perform, and the experimenter will not know which choice was made (ex : respond
with speed and accuracy)
● Vague instructions (e.g., instructions telling the participants to imagine, guess, or
visualize something) allow the participants to place their own interpretations on the
task.

Record Keeping : A checklist

● It is essential to keep good records when you conduct an experiment.
● Maintain a written sequence of conditions before the study begins and test participants
based on the random allocation to the experimental condition.
● Maintain a record of basic demographic information; the date, time, and place of
testing; and the name of the experimenter who did the testing.
● Maintain a place for the experimenter to write down comments about unusual
occurrences (e.g., a confused or uncooperative participant) or questions that come up.

Debriefing or post experimental interview

● The experiment does not—or should not—end with the completion of data collection
● Debrief with participants allows them to comment freely about any part of the
experiment.
● The interview can also provide information regarding the participants’ thinking or
strategies used during the experiment, which can help explain their behaviour