Basic Study Design

Question 1

What are the advantages of nonprobability samples?

Answer 1

They have low costs and the participants are easily accesible. Examples are the convenience and consecutive sampling method.

Question 2

What is considered the “gold standard” for generalizability?

Answer 2

Probability samples > each member of the sample has an equal chance to be part of the sample. Examples are the Simple random, systematic, stratified and cluster method.

Question 3

What are the two main goals of recruitment of participants?

Answer 3

1. Achieving a representative sample to minimize bias > nonresponse is a big problem here.
2. Recruiting enough participants to minimize the chance of getting the wrong answers

Question 4

The ultimate goal is to recruit a sample of participants that is …. (1) for the target population to control …. (2) and ….. (3) enough to control for ….. (4).

Answer 4

1. Representative
2. Systematic sources of error
3. Large
4. Random sources of error

Question 5

Explain the concepts of precision, accuracy and validity.

Answer 5

Precision = the reliability and reproducability of a measurement; determinant of power and effect size.
Accuracy = the degree to which the measure quantatively approaches the gold standard 
Validity = the degree to which the measuremnt qualitatively represents the phenomena it is intended to measure

Question 6

What is the advantage of continuous over categorical variables? And which hierarcy do we use in psychological research?

Answer 6

With continuous variables you can capture more information witch higher power and thereby need smaller sample sizes.

1. Continuous
2. Discrete
3. Ordinal
4. Nominal/dichotomous

Question 7

Which forms of variability influence the precision and accuracy of measurements?

Answer 7

1. Observer
2. Subject
3. Instrument

Question 8

How can we increase precision and accuracy of measurements?

Answer 8

Operationalization, standardizing measures, training observers, repeating and consequently summarizing these repetitions. Specifically for accuracy > blinding, calibration and unobtrusive measures.

Question 9

Explain the concepts of sensitivity and specifity

Answer 9

Sensitivity = the degree to which a measurment is able to detect differences in the variables of importance.
Specifity = the degree to which the measure only represents the variable of interest (bc often there is interference with other variables).

Question 10

What is the problem with the group study approach in brain pathology?

Answer 10

In comparing healthy controls to participants with brain pathology; differences between these are often attributed to the mere presence or absence of brain pathology. This is too simplistic and related to the modularity approach to cognition (when one function is impaired, all other functions are intact). This is a big problem and fallacy in studying brain pathology.

Question 11

Which 3 techniques can be used to decrease the methodological fallacy in NP-research (comparing HC’s to pathology group)?

Answer 11

1. Being very cautious in attributing differences between the two groups to the mere presence/absence of brain pathology, especially when the matter is underinvestigated.
2. Matching HC’s and pathology group on all variables that may affect performance
3. Performing a control task, similar to the experimental one (difficulty, setting, stimuli), but that require different cognitive functions than that are studied.

Question 12

Which four obstacles keep neuropsychologists hesitating from using computerized tests?

Answer 12

1. Psychometric; not the same psychometric properties as p+p tests, different reliability and norms; new research neccesary
2. Theoretical; low knowledge on the theoretical evidence for computerized tests (especially on validity)
3. Strategic; lack of meta-analyses while more and more tests are being developed, while they don’t know which one is the most effective > chance on splintered/fragmented profession
4. Technical; need for software updates, less flexible (you can’t explain questions, give second chances), does not take cultural/ethnic/age differences into account

Question 13

When should you use Bonferonni adjustments?

Answer 13

When you use multiple comparisons; with Bonferonni, you correct for type-I-erros (alpha). The results of a given comparison depends on how many other tests were perfomed.

Question 14

What is the goal of Bonferonni adjustments according to the Pearson-Newman theory?

Answer 14

To aid in decision making regarding repeated testing, not to assess data, as is often done now. You cannot use Bonferonni in assessing specific hypotheses! You can use it when:

1. H0 is of interest (healthy pp undergo different health checks as part of a general health check)
2. you perform repeated tests in many subsamples, without an a priori hypothesis that the primary association would differ between these groups
3. Sequential testing of trial results
4. Searching for significant associations without pre-established hypotheses (so, exploratory)

Question 15

What are examples of pitfalls in significance testing?

Answer 15

1. Post hoc testing can only be used exploratory; replication and further research is necessary to draw strong conclusions. A priori hypotheses can be tested with significance.
2. You can’t assess group differences when you assign pp randomly in your experimental research, because you use randomisation to minimize group differences.
3. Cherry picking; changing the outcome measure along the way to find significant results.
4. Only using nonspecific hypotheses > often lots of tests done, which leads to problems with the alpha adjustment > better use stepwise analysis.
5. You should not assess one hypothesis with multiple tests; then you need to correct for the alpha inflation with Bonferonni.