Statistical Psychology
Statistical Psychology
Statistical Psychology
In error
A sample
Random
If you have a child, the child can be either a male or a female. Though your second child can be either
male or female and doesn’t depend on the first child’s gender. The former is known as ____ and the
latter as _____
If you have a set of score that are too low for your liking and you decide to multiply each score by a
constant of 3, what must you do to the variance?
Probabilities require events to be Independent, but in the case of additive laws, they also require events
to be:
Disjunctive
Conjunctive
Mutually Exclusive
Rejective
All of the above
(A + B) / B
(A – B) / B
(A x B) / A
(A x B) – B
A population has a mu = 13.4, N = 550, and a variance of 64. You draw a sample of 16 from the
population and discover it has a mean of 14.2 and you end up with a variance of 55. What is the
Standard Error?
2.026
2.0000
0.571
2.345
1.854
Mu = y_bar
Mu – y_bar = 0
Mu * y_bar = y_bar^2
Mu + y_bar = mu * 2
You conduct a study where your mean attains a z-score of 1.22 on the comparison distribution, you used
a two tailed test (p < 0.05). What do you conclude?
I asked you to rate your favorite movies from most favorite as #1 all the way to #5. What type of variable
am I using?
Nominal
Ordinal
Equal-interval
Ratio
Here is a frequency table showing score on the left and frequency on the right: 0 | 0 1 | 2 2 | 4 3 | 6 4 |
8 5 | 5 6 | 2 7 | 2 8 | 1 9 | 0 What is the a) mean b) standard deviation
a) 3.967 b) 2.667
a) 3 b) 2.667
a) 4.5 b) 1.546
a) 3.967 b) 1.546
a) 3 b) 1.546
There is a controversy in using the mean to describe research. Many psychologists have proposed to:
Use mode as a way to record frequencies, then come up with a mean of frequencies to explain
populations
Use qualitative methods to obtain data, followed by quantitative methods to explain populations
Use qualitative data to describe populations and quantitative data to measure populations
The mean tends to be influenced by all sorts of scores, so there is a limit in how representative it can be
of all the scores
Central limits theorem comprises of Standard Deviation, the Mean, and the Variance, These three make
up Central Limits Theorem as the core variables to describe populations or samples
In a truly natural and random environment, variables will assume normality in a distribution
A philosophical branch comprised of those who criticize the ability of Mean, Median, and Mode in their
ability to describe statistics
Even large samples of studies do not show normality, so it is mathematically unethical to compare them
with a normal curve
Samples are usually not chosen randomly from the population (haphazard sampling)
We’re testing samples as if they were unaffected by the experimental procedure (testing as if Null
Hypothesis is true)
Please put the 5 steps of hypothesis testing in chronological order: a) determine cutoff comparison
distribution b) restate question as research hypothesis and null hypothesis c) determine sample score on
comparison distribution d) determine characteristics of comparison distribution e) decide whether to
reject the null hypothesis or not
B, d, c, e, a
B, d, c, a, e
D, c, b, a, e
B, d, a, c, e
A, b, c, d, e
The 2009 APA (6th ed.) now requires research papers to include:
Effect Size
Standard Error
If you have a population mean, and you take samples of the population many times, having a set of
sample means. When you get the mean of all the sample means, it will equal:
When figuring out the z score of a mean compared to the distribution of means, you minus the mean of
the distribution of means FROM the mean you are trying to find the z-score of, and then divide over
what?
Marginal significance occurs when a z-score comes very close to the cutoff z-score (such as z = 1.94 and
the cutoff is 1.96). Psychologists and statisticians generally:
Like this because they understand that real life is not so black/white with respect to significance
Like this because they understand there are so many guesses and estimates in statistics that it’s hard to
tell if the significant level really makes a difference
You have done a study and just calculated the confidence interval of your sample distribution (using
SEM because there are two population means). You compare your confidence interval with the general
population (comparison distribution) and notice your confidence interval does not contain the
population’s mean (null hypothesis). With this, what can you conclude?
Reject of the null hypothesis
You’ve employed a study technique to help students get a better grade on the PSYB07 midterm. You
randomly select 22 students to participate. After the midterm, the class average was 64 with a standard
deviation of 9. Your study group scores an average of 67 with a standard deviation of 6. What is the
effect size?
0.50
0.33
1.56
2.35
3.00
Increasing which of the following does not, in turn, increase power, but lower it?
Effect size
Sample size
0.20
0.50
0.80
What can you do to artificially raise power, but what are the consequences?
A more strict significance level, however this raises the probability of a Type II error
A more lenient significance level, however this increases the probability of a Type II error
A more strict significance level, however this increases the probability of a Type I error
A more leneint significant level, however this increases the probability of a Type I error
Which area of psychology research are you most likely to find a higher emphasis on the use of Effect
Sizes in research articles?
Descriptive Research
Inferential Research
Theoretical Research
Applied Research
6.5
2.167
8.667
6.333
Same as Type I
1 – alpha
1 – beta
Beta alone
A new disease called Tuition Sickness (TS) has appeared in the population. The probability of TS is 0.001.
If you have TS, you have a 0.99 probability of testing positive. However there is a 0.0002 chance of
testing a false positive. You go to your school’s Health and Wellness center to be tested for TS and it
turns out positive. What are the chances you ACTUALLY have TS given you just tested positive?
0.832
0.990
0.002
0.001
A school has 200 students (pretty small, isn’t it). 70% are men. 50% are women. 50% are blond, 50% are
brunette. What is the probability to select a blond male?
0.25
0.35
0.40
0.70
A more abstract question to challenge you. In a group you can either be A or B, and you can also be
either C or D. 80% are B’s. 40% of the population are also C’s. What is the chance of finding an A whose
also a C?
48%
32%
12%
8%
You’re conducting a study. You want to have a large effect size (0.80). In your study you know there is a
population mean of 56 and a standard deviation of 8. So you draw a sample of 10 for your experimental
procedure and assume to reject the null hypothesis on a two tailed distribution at a p
Power = 71.57%
Power = 83.22%
Power = 72.97%
Power = 80.00%
Power = 91.08%