Lecture 4

Question 1

What is the priming effect

Answer 1

Priming effect : exposure to one stimulus influences how a person responds to a subsequent stimulus
E.g. participants for whom an elderly stereotype was primed walked more slowly down the hallway when leaving the experiment than did control participants

Question 2

Ego depletion

Answer 2

Ego depletion is a psychological theory that self-control and willpower are limited and can be depleted

Question 3

Does the failure to replicate the effects from priming studies, power posing, and ego depletion mean that these effects aren’t true?

Answer 3

This does not say by itself these effects are nonexistent. Failure to replicate the study with larger sample sizes provides an evidence that these effects could have been enlarged than what it actually is, including the possibility of having non-zero effects

Question 4

In replication studies, p values were higher and effect sizes were lower. Why?

Answer 4

Since the effect size is a population characteristic, it should stay consistent in many studies despite the sample size. However, a small effect size may cause a small sample to not find significant results, resulting in it not being reported. However, this small effect size would be found by larger samples.
As a result, as the sample size increases, the average number of effect sizes decreases, as the small effect sizes are now being found and reported. Since the smaller samples for the small effect sizes are not being reported, this creates a negative relationship between sample size and effect size for published research.

Question 5

What is Publication bias?

Answer 5

researchers become interested in publishing papers with statistically significant results only. When researchers do not report null findings and treat as if it was not included in their research purpose
Evidence: Effect size does not depend on sample size. If we collected all studies’s effect sizes and corresponding sample sizes, the line should look more or less “flat” (horizontal). But there is a strong association based on published papers

Question 6

HARKing

Answer 6

HARKing (post-hoc theorizing): hypothesizing after results are known; claiming that the results of exploratory analyses were hypothesized a priori.

Question 7

Data peeking

Answer 7

Data peeking: Running preliminary analyses before data collection is complete and using results to decide whether to keep collecting data (a.k.a. ‘optional stopping’)

Question 8

p hacking

Answer 8

p hacking: Iterating through slight alterations of an analysis until p < 0.05.

Question 9

What are the 4 factors that lower the replicability of research?

Answer 9

Factors that lower replicability of research
* Publication bias : do not publish papers that are ‘statistically insignificant’.
* Data manipulation : manually change or make up entries of the dataset
* HARKing : test many different hypotheses and choose the significant result as if it was the pre-set research question
* data peeking : For the same hypothesis, try different tests or include/exclude observation manually until p < 0.05.

All these efforts lead to ‘inflated Type 1 errors’, where p values are exceedingly concentrated near 0 when the null hypothesis is true