click below
click below
Normal Size Small Size show me how
NCE Research & Pro
research & program evaluation
Question | Answer |
---|---|
experimental research is the | process of gathering data in order to make evaluative comparisons regarding different situations |
quasi-experiment | researchers use preexisting groups, IV cannot be altered (gender, ethnicity) |
internal validity | refers to whether DVs were truly influenced by the experimental IVs or whether other factors had an impact |
external validity refers to whether | the experimental research results can be generalized to larger populations (other ppl, settings or conditions) |
chi-square | nonparametric test that tests whether a distribution differs significantly from an expected theoretical distribution |
experiments emphasize parsimony - | interpreting results in simplest way |
Occam's Razor suggests that experimenters | interpret results in simplest manner |
bubbles | flaws in research (hint - sticker on car window or tint) |
undesirable variables do what to the experiment | flaw the experiment |
all correlational research is said to be | confounded or flawed |
American Psychological Association's Journal of Counseling Psychology publishes more | counseling research articles than any other periodical in our field |
basic research is conducted to advance our | understanding of theory |
applied research (action research or experience-near research) is conducted to advance our | knowledge of how theories, skills, & techniques can be used in terms of practical application |
a variable is a factor that | varies or is capable of change |
IV is a variable that the researcher | manipulates, controls, alters, or wishes to experiment with. (I manipulate the IV) |
DV is the variable the expresses the | outcome or data (Data or DV) DV expresses data regarding factors you wish to measure |
true experiment | biofeedback |
IVs & DVs can be discrete or continuous | discrete - brand of counseling or occupation; continuous - height or weight |
code of ethics for experiments | 1-subjects are informed of risks; 2-negative after effects are removed; 3-subjects can w/draw at any time; 4-confidentiality of subjects will be protected; 5-research reports results will be presented in accurate format; 6-use only techniques trained in |
control group does not receive | IV |
basic presupposition is that averages (or means) of groups do not differ significantly at | the beginning of experiment |
if you cannot ___ assign the subject to the 2 groups then your exam will consider the research a quasi-experiment | randonmly |
organismic IV is one the researcher cannot | control yet it exists (height, weight, gender) |
R. A. Fisher is known for | hypothesis testing |
hypothesis is a statement which can be ___ regarding the relationship of the IV & DV | tested |
null hypothesis suggests that there will not be a significant difference between | experimental groups which received IV and control group which did not receive IV |
in null hypothesis the IV does | not affect the DV |
meta-analysis is the study which analyzes the findings | of numerous studies |
alternative or affirmative hypothesis asserts that the IV has indeed caused a | change |
from a purely statistical standpoint, in order to compare a control group to the experimental group you will need a | test of significance |
descriptive statistics | describes data (mean, median, mode) |
P= | probability or level of significance |
P is set at | .05 or lower (.01 or .001) indicates that differences will occur by chance only 5x in 100 (significance must be set before experiment begins) |
ethnographic research is research collected by | interviews, observations, and inspection of documents |
level of significance could be referred to as | level of confidence or confidence level |
P = .05 also means | differences truly exist; experimenter will obtain the same results 95 out of 100 times |
the smaller the P value the best chance to rule out chance factors | .001 = 1 chance in 1000 or .05 = 1 in 20; .01= 1 in 100 |
Type I & Type II errors are called | alpha and beta |
Type I or alpha errors are | when researcher rejects the null hypothesis when its true |
Type II or beta errors are when researcher | accepts the null hypothesis when its false |
Type I or R = | rejects |
Type II or A = | accepts |
probability of committing a Type I error equals the level of significance; therefore, level of significance is often referred to as | alpha level |
1 minus beta is called | power of a statistical test; power connotes a statistical test's ability to reject correctly a false hypothesis |
parametric tests have more power than nonparametric tests because | parametric tests are ONLY used w/ interval and ratio data |
ALL statistical tests rely on probability because there is a | chance that results are merely chance occurrences |
researchers call these chance factors | errors |
lowering the significance level (.01 or .001) lowers Type I error but it raises the risk of committing a | Type II error |
safest bet is to set alpha at very stringent level and then use a ___ sample size, so the correct decision (to accept or reject) can be made majority of time | large |
experimenter sets the ___ level | alpha |
raising sample size ___ risks of chance/error factors | lowers |
t-test is a | simplistic form of analysis of variance |
t-test is used to see if 2 sample means are significantly different; researcher sets level of significance & then runs experiment. t-test is computed & yields a t-value. | researcher goes to t-table. if the t-value is lower than critical t - in table; accept the null hypothesis. computation must exceed # in table to reject null |
one-way ANOVA or analysis of variance (F statistic) used when there are | 2 or more groups. researcher consults an F table for critical value. If F value obtained exceeds critical F; null is rejected |
analysis of covariance or ANCOVA/ANACOVA tests 2 or more groups while | controlling for extraneous variables often called covariables |
Kruskal-Wallis is used instead of a one-way | ANOVA when data are nonparametric |
Wilcoxon -signed rank test used in place of | t-test when data are nonparametric & you wish to test whether 2 correlated means differ significantly |
Mann-Whitney U-test used to determine | whether 2 uncorrelated means differ significantly when data are non parametric |
Spearman correlation or Kendall's tau used in place of | Pearson r when parametric assumptions cannot be used |
chi-square non parametric test examines whether obtained frequencies | differ significantly from expected frequencies |
statisticians have created nonparametric tests that parallel popular | parametric tests |
when a study has more than 1 DV the term ___ of variance is used | multivariate analysis (MANOVA) |
2 IVs requires a 2 way ANOVA; 3 IVs requires a | 3-way ANOVA |
statistic that indicates the degree or magnitude of relationship between 2 variables is known as "correlation coefficient" abbreviated | r |
coefficient correlation makes a statement regarding the association of 2 variables & how a change in 1 is related to a change in another | correlations range from 0.00=no relation; 1.0 or -1.0=perfect relation |
a positive correlation is not stronger relation than negative one of same numerical value (i.e. -.70 is stronger than +.60) the minus sign just describes | the fact that as 1 variable goes up the other goes down |
positive correlation is evident when both variables change in ___ direction | same |
negative correlation is evident when variables are inversely associated - | 1 goes up the other goes down (i.e. brushing teeth more causes less cavities) |
biserial correlation is 1 variable is continuous (using interval scale) while the other is | dichotomous (i.e. correlate state licensing exam scores to NCC status (dichotomy is licensed/unlicensed) |
correlations are rarely | 1.00 (perfect) |
when 2 variables vary together this is called | covary positively |
when 1 variable increases while other decreases they are called | covary negatively |
correlational research is quasi-experimental and does not yield ___ data | cause-effect |
correlational data that describes the nature of 2 variables is called | bivariable |
if more than 2 variables are under scrutiny is it | multivariable |
N=1; N or number of persons being studied is 1. this is a "case study" of 1 approach. popular w/ behaviorists who seek overt (measurable) behavioral changes. | ex. client's dysfunctional behavior is measured (baseline measure); treatment is implemented; then behavior is measured again (i.e. another baseline is computed) As=baseline, Bs=intervention implementation, Cs=2nd or alternative form of intervention |
single case investigations are called | idiographic studies or single subject designs |
single-blind study - subject would ___ know if they are part of experimental or control group | not |
participant observer model - researcher ___ in study; while making observations about what transpired | participates |
double-blind study - experimenter and subjects are ___ of subjects' status | unaware |
experimenter effects can flaw experiment because experimenter might | unconsciously communicate intent |
AB or ABA time series design is the ___ type of single-subject research; popularized by behavior modifiers in 1960s/70s | simplest |
AB or ABA models rely on continuous measurement | i.e. baseline is secured (A); intervention is implemented (B); outcome is examined by new baseline (ABA); in order to improve research process, an ABAB design can be used |
If the pattern for 2nd AB administration mimics that of 1st AB, then chances increase that B (intervention or so-called treatment) ___ the changes rather than extraneous variable | caused |
ABA may be called "withdrawal design" or | reversal of treatment |
when researcher employs more than one target behavior it is | multiple baseline design |
correlational coefficient is a descriptive statistic which indicates degree of ___ relationship between 2 variables; also to show perfect relationship exists | linear ; (i.e. 1.0 or -1.0) graphed - a straight line is formed |
Pearson Product Moment correlation r is used for ___ or ___ data | interval ; ratio |
Spearman rho is used for ___ data | ordinal |
normal curve or Gaussian curve looks like a | symmetrical bell |
curves not symmetrical are | asymmetrical - called skewed distributions |
68-95-99.7 rule or empirical rule in normal distribution is | 68% of scores will be +/- 1 SD of mean; 95% of scores will be +/- 2 SD of mean; 99.7% scores will be +/- 3 SD of mean |
almost all scores will fall within __ SD of mean | 3 |
mode, mean, median are | most common measures of central tendency |
mode is | highest or max point of concentration; most frequently occurring score & least important measure |
median is the | middle score of distribution of scores |
mean is the | average |
modal category is the | highest value |
bimodal curve looks like | camel w/ 2 humps (i.e. men's & women's weights) |
curve w/ more than 2 humps | multimodal |
range is the distance between largest & smallest scores. To compute range | take largest number and subtract smallest # = R (sometimes adding a 1) |
larger the range the greater the | dispersion or spread of scores from the mean |
most useful measure of central tendency is the | mean or X w/ line over it |
factorial design is used when there are | 2 or more IVs |
harmonic mean refers to a central tendency statistic that is | reciprocal of the arithmetic mean of the reciprocals of the set of values |
harmonic mean cannot be used w/ | negative numbers or data containing zero |
mean is misleading when distribution is | skewed or there are extreme scores |
skewed distribution | left or right side of curve are not mirror images & mean, median, mode fall at different points |
normal curve, numbers fall | at same point |
w/ extreme scores the median is | best statistic |
median is directly in middle of numbers ranked from lowest to highest of odd numbers. If there are even numbers the median is computed by | the mean of the 2 numbers in the middle |
in factorial experiment several experimental variables are investigated & interactions are investigated & interactions can be noted; factorial designs include | 2 or more IVs |
sometimes IVs in factorial design are called | levels |
Solomon four group design the researcher uses 2 control groups (only 1 experimental group & 1 control group are pretested) the other control group & experimental group are merely post-tested; the genius of the design is | that it lets the researcher know if results are influenced by pretesting (the 2 sets of groups can then be compared) |
positively skewed - tail is to the | right & low scores fall to left |
negatively skewed - tail is to the | left & high scores to right |
raw scores don't tell us much | without more information |
histogram | bar graph |
mesokurtic refers to peakedness of a curve. the normal Gaussian curve is said to be mesokurtic since peak is in the | middle |
x axis (abscissa) is used to plot | IV and it is on horizontal axis |
y axis (ordinate) vertical axis is used for | DV (or experimental data) |
ethological observation | observations of animals |
inclusive range | highest number -- lowest number + 1 |
exclusive range | regular range; highest # -- lowest # = R |
scattergrams or scatterplot is a pictorial diagram or graph of 2 variables being | correlated |
John Henry Effect (also called compensatory rivalry of a comparison group) is a threat to the interval validity of an experiment that occurs when subjects strive to prove that an experimental treatment | that could threaten their livelihood rally isn't all that effective |
SD if everyone scored the same, the SD would be | zero |
the greater the SD the greater is the | spread |
Z-scores are same as | SD; sometimes SD are called z-scores |
T-scores or transformed scores uses mean of | 50 w/ each SD as 10 (i.e. z-score of -1.0 would be a T-score of 40; z-score of -1.5 would T-score of 35) |
stanine score divides the distribution into ___ equal parts w/ 1 being the lowest & 9 being the highest point of curve | 9 |
CEEB (College Entrance Examination Board) or ETS (Educational Testing Service) 200 -- 800 w/ mean of 500 has a SD of | 100 |
F scores are never expressed as ___ negative | number |
Kurtosis refers to the ___ of a frequency distribution | peakedness |
platykurtic distribution is ; (plat sounds like ___) | flatter & more spread out |
a tall, thin and peaked curve is called (leaps tall buildings in a single bound) | leptokurtic |
stanine (contraction of words standard and nine) stanine scores divide the distribution into 9 equals intervals w/ stanine 1 as lowest and 9th & 9 as highest. ___ is the mean | 5 |
4 basic measurement scales (noir) | nominal, ordinal, interval, ratio |
nominal is strictly qualitative data, simplest type, it | classifies, names, labels, or identifies (street address, ph #) nominal scale has true zero point & does NOT indicate order |
parametric tests use __ & ___ data | interval and ratio |
nonparametric tests us ___ & __ data | nominal and ordinal |
ordinal scale (2nd level of measurement) puts variables in | rank order; there is no math involved; ordinal sounds like order |
interval scale has #s scaled at equal distances but has NO absolute zero point. Most tests fall in this category and you can | add & subtract but not multiply or divide here |
IQ tests provide ____ measurement | interval |
ratio scale is an interval scale w/ a true zero point. highest level of measurement | adding, subtracting, multiplying, dividing are all used. Most psychological attributes CANNOT be measured on a ratio scale |
naturalistic observations | no manipulation of either variable |
2X3 factorial design uses 2 IVs. 1st IV has 2 levels (male & female) and | 2nd IV has 3 levels |
simplest form of descriptive research is ___ which requires a questionnaire return rate of ___ to be accurate. (the return is usually around 40% | survey ; 50-75% |
survey problems include | poor construction of instrument, low return rate, fact that often subjects are not picked at random, & not representative of population |
placebo | fake |
nocebo | negative effect (i.e. you may only have 2 months to live) |
Hawthorne Effect (1924-1932) | subjects perform better when they know they are being observed (reactive effect or observer) |
Rosenthal Effect | or the experimenter expectancy effect |
Observer bias | observer has perceptions regarding research that are not accurate |
Halo effect | a trait not being observed (beauty) influences a researcher's rating on another trait |
trend analysis refers to statistical procedure performed at different times to see if | a trend is evident |
ANCOVA OR ANACOVA | similar to ANOVA but more powerful b/c it can help eliminate differences between groups which otherwise could not be solely attributed to the experimental IVs |
covariate | correlates the DV |
cohort | share similarities (i.e. birth years, in same war) |
statistical regression | predicts very high and very low scores will move toward the mean if a test is given again |
quartile or fourths | 25th percentile is 1st quartile, 50th is 2nd quartile, 75th is 3rd quartile, and distance between |
cross-sectional method may be called "synchronic method" & longitudinal as | "diachronic method" |
cross sectional data are indicative of measurements or observations at | single point in time |
longitudinal study is data collected at | different points in time |
confederate or stooge | researcher has an accomplice pose as a client |
ipsative implies a within person analysis rather than a normative analysis between individuals | (i.e whether you are looking at individual's own patterns revealed by measurement (i.e. high/lows) or whether his/her score is compared to others evaluated by same measure |
chi-square - most popular | non parametric statistical test; distribution is not normal |
chi-square is merely used to determine whether an obtained distribution significantly from an | expected distribution |
demand characteristics relates to any bit of knowledge (correct or incorrect) that the subject is | aware of that can influence behavior |
summative evaluation is used to assess a | final product; summative assess how well goal has been met |
formative is ___ while program is underway | ongoing |
2 tailed test is often called a | nondirectional experimental hypothesis |
1 tailed test is a | directional experimental hypothesis |
1 tailed test have the advantage of having more power than | 2 tailed tests |
Pygmalion effect (Rosenthal/experimenter effect) experimenter falls in love w/ his own | hypothesis and experiment becomes self-fulfilling prophecy |
counterbalancing used to control for the fact the ___ of the an experiment could impact upon its outcome | order |
ahistoric therapy - any psychotherapeutic model that focuses on | here & now rather than past |
multiple treatment interference - subject receives more than__ treatment; it is tough to discern which modality truly caused the experiment | one |
SPSS - | statistical package for the social sciences - computer software program that can ease pain of computing statistical data by hand |
random (standard) sampling is | random |
stratified (stratum or strata) sampling - __ characteristic needs to be represented in sample | special |
cluster sampling used when it is nearly impossible to find a list of the ___ population | entire |
horizontal sampling occurs when researcher selects subjects from a ___ socioeconomic group | single |
vertical sampling when researcher selects persons from __ __ socioeconomic classes | 2 or more |
systematic sampling (vs. random sampling) used by taking ever nth person | (i.e. 10,000 ppl - you want 1,000 for your study, so you pick the 1st person between 1-10 at random & then use every 10th person) |
operational definition means to outline the ___ or gives details on how a test was performed | procedures |
nonparametric tests could be called | distribution free tests |
matched design subjects are matched in regard to any variable that could be | correlated w/ DV |
Mann-Whitney U-test used to determine whether 2 | uncorrelated/unmatched means differ significantly |
Wilcoxon signed-ranked test examines whether 2 | correlated means differ significantly |
Wilcoxon is an alternative to | t-tests when parametric precepts cannot be accepted |
deductive reduces the | general to the specific |
inductive goes from | specific to general |
attrition or experimental mortality subjects that | drop out of study |
standard error of measurement (SEM) tells counselor what would most likely occur if the same individual took | same test again |