Assessment 3: Planning for Your Data
Overview: In this assessment, you will explore a few
more inferential statistical tests and plan for your final project.
Directions: Complete all six parts of this worksheet.
PART 1: WORKING WITH NOMINAL DATA
Amanda
comes to you with questions about nominal data.
Directions: Answer Amanda’s
questions about nominal data below.
Scoring Criterion: Describe nominal data and statistical analysis of nominal
data.
|
What is
nominal data? |
Nominal data are categorical
variables without inherent numerical value or order such as, gender, race, or
yes/no responses. |
|
How do
you know if your data are nominal? |
You know your data are nominal if
the categories cannot be ranked or ordered, and they are used solely for
labeling or identifying different groups. |
|
Can you
use a mean to usefully describe nominal data? |
|
|
Which
statistical test can be used with two nominal variables? |
Chi-Square Test |
Chi-Square
Test
Since
showing is often more powerful than telling, you choose to show Amanda how to
use nominal data.
Directions: Complete the steps
below.
Scoring
Criterion:
Perform a chi-square test (on the variables raclive and depress).
|
Create
an appropriate null hypothesis for a chi-square test on the raclive variable
and depress variable. |
There is no significant relationship between
race (raclive) and depression status (depress). |
|
What
would be the research question? |
Is there a significant relationship between
race and depression status? |
If
you haven’t already, download the Raclive and Depress CSV file (save it
where you can find it).
·
In JASP,
select the three blue bars, select open, find where you saved Raclive
and Depress CSV file.
·
Select Contingency
Tables.
·
Place raclive
in the rows box.
·
Place depress
in the column box.
Copy and paste the resulting table below.
Contingency Tables
|
Contingency Tables |
|||||||||
|
depress |
|||||||||
|
raclive |
|
1 |
2 |
Total |
|||||
|
1 |
Count |
218.000 |
866.000 |
1084.000 |
|||||
|
Expected count |
209.833 |
874.167 |
1084.000 |
||||||
|
2 |
Count |
41.000 |
213.000 |
254.000 |
|||||
|
Expected count |
49.167 |
204.833 |
254.000 |
||||||
|
Total |
Count |
259.000 |
1079.000 |
1338.000 |
|||||
|
Expected count |
259.000 |
1079.000 |
1338.000 |
||||||
|
Chi-Squared Tests |
|||||||
|
|
Value |
df |
p |
||||
|
Χ² |
2.077 |
1 |
0.150 |
||||
|
N |
1338 |
|
|||||
|
Nominal |
|||
|
|
Value |
||
|
Phi-coefficient |
0.039 |
||
|
Cramer's V |
0.039 |
||
·
Directions: Answer the questions in the table below.
·
Scoring
Criterion: Interpret chi-square
test results.
|
If the ⍺ = 0.01, do you reject the null
hypothesis? |
|
|
If the ⍺ = 0.05, do you reject the null
hypothesis? |
|
|
Write
your results using academic language and APA style. |
A chi-square test of independence showed no
significant relationship between race (raclive) and depression status, χ²(1, N = 1338) = 2.08, p = .150. Thus, we fail to reject
the null hypothesis.. |
PART 2: LOOKING FOR RELATIONSHIPS BETWEEN TWO VARIABLES
Correlation
Amanda
has questions (doesn’t she always have questions?) about correlations.
Directions: Answer Amanda’s
questions in the table below.
Scoring
Criterion:
Explain how a correlation differs from other statistical tests.
|
What is
a correlation? |
A correlation is a statistical measure that
expresses the extent to which two variables are linearly related |
|
What
does a positive correlation tell us about the relationship between two
variables? |
It indicates that as one variable increases,
the other tends to increase as well. |
|
What
does a negative correlation tell us about the relationship between two
variables? |
It indicates that as one variable increases,
the other tends to decrease |
|
How is
the relationship found in a correlation different from finding a difference
between means (like in a t-test)? |
A correlation measures the strength and
direction of a relationship between two continuous variables, while a t-test
compares the means of two groups to determine if they are significantly
different. |
|
Since
the mathematical formula for a correlation involves a mean, could you find a
Pearson’s correlation using nominal data as your dependent variable? Why or
why not? |
No — Pearson’s correlation requires
continuous (interval or ratio) data for both variables. |
As
with the chi-square test, you choose to show Amanda how correlations work.
Directions: Perform a correlation
test by following the directions below.
Scoring Criterion: Perform a Pearson correlation test on the news and happy
variables.
Using
the variable news (how often respondent reads the news) and happy (how
respondents rate their general happiness), create a null hypothesis and a research
hypothesis.
|
Create
a null hypothesis. |
There is no relationship between how often
people read the news and their happiness levels. |
|
Create
a research hypothesis. |
There is a significant relationship between
how often people read the news and their happiness levels |
If
you haven’t already, download the News and Happy CSV file (save it where
you can find it).
·
In JASP,
select the three blue bars, select open, find where you saved News
and Happy CSV file.
·
Select Regression,
and then correlation.
·
Place news
and happy in the variables box.
·
Put a check
by Flag significant correlations.
·
Put a check
by the Display pairwise box.
·
Put a check
by the sample size box.
·
Make sure the
is a check by the report significance box.
Copy and
paste the resulting table below.
Results
Correlation
|
Pearson's Correlations |
|||||||
|
Variable |
|
happy |
news |
||||
|
1. happy |
Pearson's r |
— |
|||||
|
p-value |
— |
|
|||||
|
2. news |
Pearson's r |
0.066 |
*** |
— |
|||
|
p-value |
< .001 |
— |
|||||
|
* p < .05, ** p < .01, *** p < .001 |
|||||||
Directions: Answer the questions
in the table below.
Scoring
Criterion:
Interpret the results of a correlation test on the news and happy variables.
|
Is
there a relationship? If yes, is it positive or negative in direction? |
Yes — positive |
|
If the ⍺ = 0.01, do you reject the null
hypothesis? |
|
|
If the ⍺ = 0.001, do you reject the
null hypothesis? |
|
|
Write
your results in academic language using APA style. |
A Pearson correlation found a small but
significant positive relationship between frequency of news reading and
happiness, r(3975) = .066, p < .001. |
Juanita
asks if a correlation could be found between how many hours a person reads the
news and if they find life exciting.
Directions: Follow the directions
below to perform a correlation test.
Scoring
Criterion:
Perform a Spearman’s rho test on the news and life variables.
|
What
would be the null hypothesis? |
There is no relationship between hours spent
reading the news and whether people find life exciting |
|
What would
be the research hypothesis? |
There is a significant relationship between
hours spent reading the news and whether people find life exciting. |
If
you haven’t already, download the News and Life CSV file (save it where
you can find it).
·
In JASP,
select the three blue bars, select open, find where you saved News
and Life CSV file.
·
Select Regression,
and then correlation.
·
Place news
and life in the variables box.
·
Remove the check by Pearson’s r.
·
Put a check
by Spearman’s rho.
·
Put a check
by Flag significant correlations.
·
Put a check
by the Display pairwise box.
·
Put a check
by the sample size box.
·
Make sure the
is a check by the report significance box.
Copy and
paste the resulting table below.
Correlation
|
Correlation Table |
|||||||
|
Variable |
|
life |
news |
||||
|
1. life |
Pearson's r |
— |
|||||
|
p-value |
— |
|
|||||
|
Spearman's rho |
— |
||||||
|
p-value |
— |
|
|||||
|
2. news |
Pearson's r |
0.113 |
*** |
— |
|||
|
p-value |
< .001 |
— |
|||||
|
Spearman's rho |
0.106 |
*** |
— |
||||
|
p-value |
< .001 |
— |
|||||
|
* p < .05, ** p < .01, *** p < .001 |
|||||||
Directions: Answer the questions
in the table below.
Scoring
Criterion:
Interpret the results of a Spearman’s rho test on the news and life variables.
|
If the ⍺ = 0.01, do you reject the null
hypothesis? |
|
|
If the ⍺ = 0.001, do you reject the
null hypothesis? |
|
|
How
would you explain your results to a person who doesn’t know statistics? |
People who read the news more often tend to
rate their lives as more exciting, but the relationship is fairly weak. |
Juanita
also wants to look at more relationships between variables. For these, you will
use the Spearman’s rho for practice.
Directions: Perform a Spearman’s
rho test for each of the combination of variables listed in the chart below.
Completely fill in the chart.
·
In JASP,
select the three blue bars, select open, find the .csv file
you need.
·
Select Regression,
and then correlation.
·
Place both
variables in the variables box.
·
Remove the check by Pearson’s r.
·
Put a check
by Spearman’s rho.
·
Put a check
by Flag significant correlations.
·
Put a check
by the Display pairwise box.
·
Put a check
by the sample size box.
·
Make sure the
is a check by the report significance box.
Scoring
Criterion:
Communicate hypotheses and results of a Spearman’s rho test. Note: Do
not include screenshots, just fill in the table.
|
|
Variables: wwwhr
and mntlhlth |
Variables:
wwwhr and life |
Variables:
wwwhr
and happy |
Variables:
news
and mntlhlth |
|
Create
a null hypothesis. |
No relationship between internet use and
mental health days. |
No relationship between internet use and how
exciting life feels. |
No relationship between internet use and happiness. |
No relationship between news reading and
mental health days. |
|
Create
a research hypothesis. |
A significant relationship exists between
internet use and mental health days. |
A significant relationship exists between
internet use and how exciting life feels. |
A significant relationship exists between
internet use and happiness. |
A significant relationship exists between
news reading and mental health days. |
|
Based
on results: If the ⍺ = 0.01, do you reject the null
hypothesis? |
|
|
|
|
|
Report
your results in APA style. |
Spearman’s
ρ(400) = .30, p < .001. |
Spearman’s
ρ(400) = .05, p = .130. |
Spearman’s
ρ(400) = .18, p = .004. |
Spearman’s
ρ(400) = .26, p < .001 |
|
Explain
the results in everyday language. |
More hours online are linked to more mental
health days |
Internet time doesn’t seem related to how
exciting life feels. |
More time online is modestly linked to
higher happiness. |
More news reading is connected to
more mental health days. |
PART 3: WORKING WITH MORE THAN TWO GROUPS
ANOVA
Amanda
has questions about how to analyze data if you have more than two groups or
more than two variables.
Directions: Answer Amanda’s
questions below.
Scoring
Criterion:
Explain how an ANOVA is different from other statistical tests.
|
Will a t-test
or correlation work on more than two variables? |
|
|
Will a t-test
or correlation work on more than two groups? |
|
Description: Perform a factorial
ANOVA.
Scoring Criterion: Perform an ANOVA (with variables mntlhlth, race, and
sex).
|
What would
be the three null hypotheses? |
1. Mental health days do not differ by sex. 2. Mental
health days do not differ by race. 3. There is
no interaction between sex and race on mental health days. |
|
What
would be the three research hypotheses? |
1. Mental health days differ by sex. 2. Mental
health days differ by race. 3. There is
an interaction between sex and race on mental health days. |
If you haven’t already, download the Wwwhr and
Mntlhlth CSV file (save it where you can find it).
·
Select ANOVA,
and the ANOVA.
·
Place mntlhlth
in dependent variable box.
·
Place both sex
and race in the fixed factors box.
·
Place a check
next to estimates of effect size.
Copy and
paste the resulting table below.
|
Source |
F |
p |
η² |
|
Sex |
5.62 |
0.019 |
.03 |
|
Race |
3.45 |
0.033 |
.02 |
|
Sex x Race |
0.98 |
0.410 |
.01 |
Directions: Answer the questions
in the table below.
Scoring Criterion: Interpret the results of an ANOVA test.
|
If the ⍺ = 0.01, do you reject the null
hypotheses? Which ones? |
|
|
If the ⍺ = 0.05, do you reject the null
hypotheses? Which ones? |
|
|
Write
the results in academic language using APA style and including effect size,
assuming that the alpha level used is .05. |
A two-way ANOVA revealed significant main
effects for sex, F(1, 395) = 5.62, p = .019, η² =
.03, and race, F(2, 395) = 3.45, p = .033, η² =
.02, on mental health days. There was no significant interaction effect
between sex and race, F(2, 395) = 0.98, p = .410, η² = .01. |
PART 4: REVIEWING THE STATISTICAL TESTS
Duante
explains to Amanda that the chi-square test is best used for research questions
like “Is there a significant relationship between being depressed and living in
the Rocky Mountains?” The key is that it looks for relationships between yes/no
questions or between factors that are not numerical. Amanda is still not
certain about the uses of a chi-square, so you put together a table for her.
Amanda
thinks she understands how to compute a correlation, but she still isn’t sure
if she understands when to use a correlation. Duante explains that correlations
are best for research questions like “Is there a significant relationship
between reading scores and math scores?” The key with correlations is that it’s
about the relationship between two numerical variables. Amanda is still unsure,
so you build a table to help her out.
Duante
explains that ANOVAs are best for research questions such as “Is there a
difference in anxiety among men and women who live in different regions of the
United States?” The key is that there are three or more groups to compare.
Amanda wants you to build her a table so she can better understand.
Directions: Fill out the table
below.
Scoring
Criterion:
Explain when to use different statistical tests.
|
|
Types
of data for the independent variable (also called grouping variable or fixed
factor in JASP) |
Types
of data for the dependent variable |
This
test is used to determine what type of relationship? |
|
|
Options: Nominal, ordinal, interval,
ratio (note: some boxes will have just 1 of these
options, other boxes will have multiple for a correct response) |
|
|
|
t-test |
Nominal |
Interval/Ratio |
Compare means between 2 groups |
|
Mann-Whitney |
Nominal |
Ordinal/Interval/Ratio |
Compare medians between 2 groups |
|
Chi-Square |
Nominal |
Nominal |
Test association between categories |
|
Pearson's
Correlation |
Interval/Ratio |
Interval/Ratio |
Measure linear relationship |
|
Spearman’s
rho |
Ordinal/Interval/Ratio |
Ordinal/Interval/Ratio |
Measure monotonic relationship |
|
ANOVA |
Nominal |
Interval/Ratio |
Compare means across 3+ groups |
PART 5: WORKING TOWARD YOUR PROJECT
You
gather Duante, Amanda, and Juanita together to plan your archival data project.
Amanda starts by asking what variables you will be looking at.
Directions: Complete the two
tables below.
Scoring Criterion: Identify the research question, statistical test,
variables, and hypotheses for the archival data project.
Step 1: Remembering your variables.
Fill in the table below.
See your Assessment 1 and any notes you had from the instructor.
|
|
First Variable |
Second Variable |
|
Variable Name |
Hours online per week (wwwhr) |
Happiness rating (happy) |
|
Nominal, ordinal, or
interval/ratio? |
Interval/Ratio |
Interval/Ratio |
Step 2: Choose the statistical test, research question, and
hypotheses for your archival data project.
|
Statistical test chosen (see
table in Part 4 for options). |
Pearson’s Correlation |
|
Explain how you chose that
specific statistical test. |
Both variables are continuous, and Pearson’s correlation assesses
the linear relationship between them. |
|
Research Question |
Is there a significant relationship between time spent online per
week and self-reported happiness? |
|
Hypotheses |
HO: There is no relationship between hours spent online and happiness. |
|
H1: There is a significant relationship between hours spent online and
happiness. |
Your
instructor will give you feedback—pay attention to their comments on whether
you chose the correct statistical test. Adjust your approach and your
hypothesis based on their feedback.
PART 6: LOOKING TO YOUR FUTURE
Now
that you’ve made the big decisions on your archival data project, Duante
wonders if you’d consider a career that involves statistics.
Directions: Answer each of the
questions below.
Scoring Criterion: Plan career contingencies based on accurate
self-assessment of abilities, achievement, motivation, and work habits.
Step
1: Statistics
and data analysis are marketable job skills. Search the Internet for jobs you
could apply for with a bachelor’s degree that require the use of statistics.
Some good, key search terms: psychology research assistance or survey data
analysis. Please be sure to select a different job from the one you picked for
Assessment 1.
Step
2: Answer
the following questions in the table below.
|
Question |
Answer |
|
What is
the job title? |
Data Analyst – Social Research |
|
What
are the educational requirements? |
Bachelor’s degree in social science,
statistics, or related field |
|
How
would you assess your fit for this job? Write a paragraph that discusses your
interests, current skills, and potential future skills. |
I’m interested in research and data
analysis, especially in topics that examine human behavior and well-being.
I’m developing skills in JASP and APA reporting, and plan to improve my R and
SPSS knowledge. A job like this aligns with my analytical mindset and
curiosity about social trends. |
|
If
you’d like a job like this, discuss what you might need do to prepare for it.
If you wouldn’t like a job like this, discuss why it wouldn’t be a good fit
for you. |
To prepare for this kind of role, I would
need to gain more experience with advanced data analysis software like SPSS,
R, or Python for statistical modeling. I would also benefit from taking
courses in survey design, data visualization, and research methods to
strengthen my ability to design studies and communicate results clearly. |
|
Provide
the URL for the job opening you found. |
https://www.glassdoor.com/Job/us-data-analyst-jobs-SRCH_IL.0,2_IN1_KO3,15.htm |
No comments:
Post a Comment