Scribd
Upload
26 views11 pages

Lecture 21 STATS 30301

Lecture 21 STATS 30301

Uploaded by

yoav.naturman123
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
26 views11 pages

Lecture 21 STATS 30301

Lecture 21 STATS 30301

Uploaded by

yoav.naturman123
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 11

Lecture 21

STAT 30301 – Probability & Statistics for


Business
We Already Know:
Inference: use confidence interval to estimate ; use hypothesis test to measure strength of evidence
1-sample Inference: 1 –sample z-test/interval about or p ; 1 – sample t-test/interval about
About : is known & X is Normal or n>30, use standard normal distribution; H0:
Standard Error Confidence Interval Test Statistic
𝑥 − 𝜇0
𝜎 x̅ 𝑧=
𝜎
√𝑛 √𝑛
About : is unknown & X is not strongly skewed/no outliers, use t-distribution with (n – 1) degrees of freedom; H0:
Standard Error Confidence Interval Test Statistic

𝑠 x̅ 𝑡=
𝑥 − 𝜇0
𝑠
√𝑛 √𝑛
About p:  approximately Normal if np 5 and n(1-p) 5; = ; H 0 : p = p0
Standard Error Confidence Interval Test Statistic

√ √
^ − 𝑝0
𝑝
𝑝 (1− 𝑝) ^ − 𝑝)
𝑝(1 ^ 𝑧=

𝑛
^ ± 𝑧 𝛼 /2
𝑝
𝑛 √ 𝑝 0 (1 − 𝑝 0 )
𝑛
Example: Local vs. Online
Prices
A local insurance agent was concerned about an advertising campaign
that claimed cheaper car insurance rates from an online company. To
investigate, they randomly selected profiles (car type, coverage,
driving record, etc.) for 10 of their clients and checked the online price
quotes for equivalent policies. Below are the prices for each:

Local 568 872 451 1229 605 1021 783 844 907 712

Online 391 602 488 903 677 1270 703 789 1008 702

Do the data provide evidence that the online company’s rates are less?
Matched Samples or Paired
Data:
• Paired data involve taking two measurements on the same
element, or one observation on each of two very similar
(identical) elements.
• Recall the conditions for inference: randomness, Normal-shaped
distribution, and independence
• When an element is measured twice, those two measurements
are necessarily dependent.
• To obtain independence, we find the difference, di, of each pair;
together, all dis produce a single sample of independent
differences.
Inference About , the Mean Difference of
Paired
Let = the population mean Data:
of the differences,
then = the sample mean of the differences
Ha:
H0 : OR
Ha: where is usually 0
Your textbook uses the sign that is OR
the opposite of Ha but it always has Ha:
an equal sign: or or just

When σ is unknown & D is not strongly skewed/no outliers, use t-distribution with (n – 1) degrees of freedom
Standard Error Confidence Interval Test Statistic
Example: Hypothesis Test
about
Profile 1 2 3 4 5 6 7 8 9 10
Local 568 872 451 1229 605 1021 783 844 907 712
Online 391 602 488 903 677 1270 703 789 1008 702

d 177 270 –37 326 –72 –249 80 55 –101 10

Then: n = 10, = 45.9, and = 175.663; df = 9; Let = 0.05


Define the parameter: (inthedollars)
population mean difference between local and online policy rates
=

for clients at this insurance agency. So D = Local - Online

Check conditions: Randomness is stated; dis are independent;


Normality: Assume the distribution of differences is not
strongly skewed and is without outliers
State the hypotheses:
Calculate the test statistic:

Find the p-value:

Find the critical value:

State the conclusions:


Since p-value (or since t.s. = 0.83 c.v = 1.833), Do not reject H0 at = 0.05.
Conclude there is not evidence that the population mean difference between local and
online policy rates (in dollars) for clients at this insurance agency is greater than zero.
Example: Math Game Test
Scores
A study is conducted to determine whether a recently released
educational game improves the mathematical abilities of students in a
certain 2nd grade class. Eight students are randomly selected, and their
math skills are given a score from 1 to 10 based on a pretest. After the
students play the educational game, they are given a posttest.

Conduct a test, at = 0.10, to determine if the mathematical test scores


of 2nd graders in this particular class improved, on average, after
playing the educational game.

Then find the 90% confidence interval.


Matched Pairs: Math Game
Test Scores
Students
Pre-game Post-game Difference:
score score (Post – Pre)
Standard deviation
1 2.56 4.54 1.98
of the differences:
2 3.22 5.33 2.11 (from difference data set)
3 3.45 4.32 0.87

4 5.55 7.45 1.9


Sd = 1.474
5 5.63 7.00 1.37

6 7.89 9.80 1.91 **Be Careful:


7 7.66 5.33 –2.33 The standard deviation of the
8 6.20 6.80 0.6
differences is NOT the difference
of standard deviations!!
Define the parameter: = the population mean difference between post-test and pre-test scores for
students in this class. So, D = post - pre
Check conditions: Randomness is stated; Normality? Assume no strong skewness or outliers;
Independence? Assume 160 students in class

State the hypotheses:

Calculate the test statistic:

Find the p-value:


Find the critical value:
State the conclusions: ______ H0 at = 0.1. Conclude there is ______ that the population mean _______
between post-test and pre-test scores for students in this class is greater than zero.
Construct the 90% Confidence Interval: 𝑡 0.05 ,7 =1.895 ; = 0.9876 ; : (______, ______)
We are 90% confident that the population mean difference between post-test and pre-test scores for students in
this class is between _______ and _______ points. NOTE: _____ is not a plausible value for the mean difference.
Example: Weight Gain
16 randomly selected non-obese adults consumed 1000 calories per day in
excess of the calories needed to maintain a stable body weight. The subjects
stayed on this diet for 8 weeks, consuming a total of 56,000 extra calories.
According to theory we would expect these subjects to gain 7.25 kg. Below are
the weights (in kg) before (B) and after (A) the 8-week study. Test, at = 0.05,
if the weight gain is less than expected.
Let d = difference of after and before in kg (A – B)
B 55.7 54.9 59.6 62.3 74.2 75.6 70.7 53.3 73.3 63.4 68.1 73.7 91.7 55.9 61.7 57.8

A 61.7 58.8 66.0 66.2 79.0 82.3 74.3 59.3 79.1 66.0 73.4 76.9 93.1 63.0 68.2 60.3

d 6 3.9 6.4 3.9 4.8 6.7 3.6 6 5.8 2.6 5.3 3.2 1.4 7.1 6.5 2.5

p-value < 0.005; 95% Conf. Interval:


Ho: critical value: t = –1.753
t = –5.77 (3.8, 5.66)
Ha: Reject H0 at = 0.05

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy