Facts for statistics 1 & 2
Lecture 1/2:
Simple random sampling: every member has an equal chance to be selected
Non-random sampling: probability of selection unknown to the researcher. sampling biases.
Parameter: numerical summary of the population –Greek letter
Statistic: numerical summary of the sample –by Latin letter
Sampling error: difference between parameter and statistic
Measurement: quantified difference on a variable
Nominal: numbers only indicate group membership. Numbers function as labels. (categorical)
Ordinal: numbers indicate ordered groups. (categorical)
Interval: numbers form a scale with an arbitrary zero-point and arbitrary unit. (continuous)
Ratio level: numbers form a scale with a non-arbitrary zero-point (or: zero has a meaning). (continuous)
Mode: the score that is observed most frequently
Median: (middle value); the score that separated the higher from the lower half (/mean of 2 middles)
Range: distance from lowest score to highest score
Variance: Sx2 = SUM OF (xi – xmean) / (N-1)
Standard deviations (SD): Sx = wortel (Sx2)
P(A|B): read as “probability of A, given B”
Z = (x – μ) / σ - 68% rule (1 SD) – 95% rule (2 SD’s)
Lecture 3/4:
Sample results/mean: sensitive to sampling fluctuations. N < 30 is reasonable.
3 types distribution: Population distribution, Distribution for a sample, Sampling distribution.
Standard deviation = standard error = σx = (σ/wortel(n))
Test statistic: (sample value – expected value under H0) / standard error.
Type I Error: rejecting H0 while it is true
Type II Error: retaining H0 while it is false
Reject H0 when observed test statistic (t) is larger or equal to critical t value (defining bounds)
Lecture 5/6:
T = (Xmean-μH0) / (Sx / wortel (N) )
σ2 is the pooled variance (for both populations)
Levene’s test not significant: assume equal variances.
Effect size: difference between the value you specified for H0, and the value of your sample statistic.
Between-subjects design: for each condition, we have an independent sample of persons
Within-subjects design: subjects may be exposed to multiple conditions
ANOVA: comparing means of groups
Score of j in group i = grand mean + group effect (deviation group i from grand mean = μi - μ) + residual
(deviation j in group i = Yij - μi) Yij = μ + ai + eij
SSbetween: SUM of ni (Yi – Y)2
SSwithin/SSresidual: SUM of SUM of (Yij – Yi)2
Lecture 1/2:
Simple random sampling: every member has an equal chance to be selected
Non-random sampling: probability of selection unknown to the researcher. sampling biases.
Parameter: numerical summary of the population –Greek letter
Statistic: numerical summary of the sample –by Latin letter
Sampling error: difference between parameter and statistic
Measurement: quantified difference on a variable
Nominal: numbers only indicate group membership. Numbers function as labels. (categorical)
Ordinal: numbers indicate ordered groups. (categorical)
Interval: numbers form a scale with an arbitrary zero-point and arbitrary unit. (continuous)
Ratio level: numbers form a scale with a non-arbitrary zero-point (or: zero has a meaning). (continuous)
Mode: the score that is observed most frequently
Median: (middle value); the score that separated the higher from the lower half (/mean of 2 middles)
Range: distance from lowest score to highest score
Variance: Sx2 = SUM OF (xi – xmean) / (N-1)
Standard deviations (SD): Sx = wortel (Sx2)
P(A|B): read as “probability of A, given B”
Z = (x – μ) / σ - 68% rule (1 SD) – 95% rule (2 SD’s)
Lecture 3/4:
Sample results/mean: sensitive to sampling fluctuations. N < 30 is reasonable.
3 types distribution: Population distribution, Distribution for a sample, Sampling distribution.
Standard deviation = standard error = σx = (σ/wortel(n))
Test statistic: (sample value – expected value under H0) / standard error.
Type I Error: rejecting H0 while it is true
Type II Error: retaining H0 while it is false
Reject H0 when observed test statistic (t) is larger or equal to critical t value (defining bounds)
Lecture 5/6:
T = (Xmean-μH0) / (Sx / wortel (N) )
σ2 is the pooled variance (for both populations)
Levene’s test not significant: assume equal variances.
Effect size: difference between the value you specified for H0, and the value of your sample statistic.
Between-subjects design: for each condition, we have an independent sample of persons
Within-subjects design: subjects may be exposed to multiple conditions
ANOVA: comparing means of groups
Score of j in group i = grand mean + group effect (deviation group i from grand mean = μi - μ) + residual
(deviation j in group i = Yij - μi) Yij = μ + ai + eij
SSbetween: SUM of ni (Yi – Y)2
SSwithin/SSresidual: SUM of SUM of (Yij – Yi)2
