When Data Lies
What data manipulation technique involves selecting only the data that supports your desired conclusion?
Cherry Picking
What type of correlation occurs when one variable increases as the other decreases?
Negative Correlation
Which statistical test is used to measure the relationship between two numerical variables?
Pearson's Correlation
What does a low standard deviation indicate about the data?
Data is clustered closely around the mean
What philosophical principle states that the simplest explanation with the fewest assumptions is usually the best?
Occam's Razor
What graphing trick exaggerates trends by making differences appear larger than they really are?
Misleading Axes
True or False: Correlation automatically proves causation.
False
Which statistical test is commonly used to compare the means of two groups?
t-Test
Approximately how many false positives are expected if 20 significance tests are conducted at p < 0.05?
One False Positive
What philosophical approach argues that knowledge comes primarily from observation and evidence?
Empiricism
What unethical statistical practice involves running many tests until a significant result appears by chance?
P-Hacking
What is a third variable that affects both the suspected cause and effect, creating a misleading relationship?
Confounding Variable
Which test determines whether observed categorical data differs from what would be expected by chance?
Chi-squared Test
What process completely removes outliers from a dataset?
Trimming
What is the name of the assumption that there is no effect, relationship, or difference between variables?
Null Hypothesis
What practice occurs when researchers only publish positive results while negative findings disappear?
Publication Bias
Name one requirement that must be met before causation can be claimed.
Cause must come first / Correlation must exist / Alternative explanations must be ruled out
Which statistical test extends the t-test to compare three or more groups?
ANOVA (F-test)
What process replaces extreme outliers with the highest non-outlier value?
To avoid p-hacking, what should researchers do before collecting data?
Pre-register their hypothesis
A researcher forms a hypothesis after seeing the results and presents it as if it was made beforehand. What is this called?
HARKing - Hypothesising after results are known
A Pearson's correlation coefficient of +0.85 indicates what type of relationship?
Strong Positive Correlation
A study reports p = 0.004. How would this level of evidence be classified?
Strong Evidence (p < 0.01)
Why can trimming and winsorising be problematic?
They distort data and may remove valuable information
To avoid HARKing, researchers should clearly distinguish between what two types of analyses?
Exploratory Analysis and Confirmatory Analysis