to reconstruct the past systematically, objectively, and accurately, often in relation to a hypothesis or theme.
to describe systematically a situation or area of interest factually and accurately. This type of research is usually not conducted for the purpose of hypothesis testing.
to investigate patterns and sequences of growth and/or change as a function of time.
Case and Field Research
to study intensively the background, current status, and environmental interactions of a given social unit: an individual, group, institution, or community. This type of study is regularly used in anthropological research, but can be used to do an information requirements analysis for an information system.
Cor relational Research
to investigate the extent to which variations in one factor correspond with variations in one or more other factors based on correlation coefficients.
Casual Comparative or Ex Post Facto
to investigate possible "cause-and-effect" relationships by observing some existing consequence and searching back through the data for plausible causal factors. Cause-and- effect relationships are very difficult to determine in this type of study. There is already a consequence so there is no direct control by the researcher.
True Experimental Research
to investigate possible cause-and-effect relationships by exposing one or more experimental groups to one or more treatment conditions and comparing the results to one or more control groups not receiving the treatment (random assignment being essential). The researcher has control of the group assignment and does so without bias (randomization). This is in contrast to ex post facto studies where the researcher has no control of the group assignment (such as readers versus non readers). Experimental research is often difficult to do in a natural setting.
to approximate the conditions of the true experiment in a setting which does not allow the control and or manipulation of all the relevant variables. This type of study is more flexible and adaptable to a natural setting than a true experiment because it does not assume assignment to groups (such as treatment and control) without bias (meaning at random).
A strategy in educational research enabling problem solving in the natural setting and depending on participatory enthusiasm, as it often results in organizational change.
Qualitative v. Quantitative
If you are observing and writing descriptions of the behavior of a group of people, that is qualitative. If you are looking at the frequencies of certain types of behavior, that is quantitative.Basically, you can think of qualitative data as words, and quantitative data as numbers.
Numbers that describe a situation or problem of interest and provide an efficient summary. Examples-- average family size literacy rates unemployment figures batting average average daily circulation percentages, averages, rates of change
Tests which we use in research to draw conclusions about groups of data. In research you make hypotheses, collect data, apply tests to this data, and make decisions/conclusions concerning the results of these tests, more powerful than descriptive statistics because you can draw conclusions; however, this function involves probability and some risk. There is always a certain probability you could be wrong in your conclusion.
Information or observations collected in order to measure or describe a situation or problem of interest.
Objects or concepts which must have a value or a definition assigned to them in order that they can be measured and analyzed.
How a variable is defined or measured.
A set of concepts (variables) plus the interrelationships that are assumed to exist among those concepts.
The consequences of our theoretical assumptions--the statements we submit to testing.
A variable which is thought to influence another variable.
A variable which is influenced or is the consequence of the independent variable.
Our hunch or assumption--what we plan to test through our research.
A statistical hypothesis--an assumption that no relationship exists.
Research - 3 Qualities defining the term
Empirical, meaning based on evidence or data-centered. 2. Systematic, meaning by use of a method. 3. Objective, meaning it is presumably conducted and interpreted by the researcher without bias.
the researcher is testing theory and ideas without necessarily applying the results to practical problems. A classic example is from learning experiments where subjects were asked to recall a series of paired-associate nonsense syllables.
Applied Research (syn.Action, Field, Evaluation)
used to influence policy and decision-making, and is conducted to solve problems (often immediate problems), sometimes only within one organization (and, results are only applicable to that organization).
A scale of measurement that involves distinct categories or classification. The categories may differ (not be equal), but you can't say one category is greater than (in terms of quantity) then another category.may use numbers as categories (such as room numbers), but the numbers do not have quantitative properties. So you would not add room 205 to room 209!
A scale of measurement that possesses an inherent order or different ranking. The rank or the scale points indicate greater than or lesser than, but we don't know the magnitude (amount) of the difference between the ranks or scale points or "how much" difference.
A scale of measurement that provides fixed units of measurement and an arbitrary zero point. An arbitrary zero means that zero does not indicate a complete absence of the characteristic being measured.characterized by equal intervals, you can determine the amount of difference between two numbers (unlike ordinal), but because you don't have a base of zero, you cannot say one number is twice as great as another. You can use addition, subtraction, etc. with this type of scale.
A level of measurements that has the same characteristics of an interval scale (fixed units of measurement/equal intervals) with the additional characteristic of an absolute zero point. Anything that can be measured from a base of zero: the only one of the four levels of measurement where you can make the statement "twice as much."
The number of times a given value of a variable occurs
A means of visually displaying data; a table that lists the individual observations for a variable and the number of times a given value occurs
Construction of the Frequency Distribution Table
variable (labeled X) and listed the number of stops from highest to lowest (lowest to highest is also an option). Beside this list, we placed the frequency (symbolized as F), or the number of times each observation/value occurs. "N" represents the number of (cases).
consists of a set of bars--the position of a bar is over the value of the variable, the height of the bar represents the frequency of the value.The frequency is always located on the vertical axis (called the Y axis), and the variable is always located on the horizontal axis (X axis).
3/4 high rule
This rule states that the vertical or Y axis should be constructed so that the height of the maximum point (the highest bar) is approximately equal to ¾ the length of the horizontal or X axis. This rule is important for providing a standard height to the bars. It is very difficult, but necessary, to create standards for constructing graphs, considering variables have different units of measure, scales, etc. Standards are to keep "well-meaning" individuals from stretching the truth with statistics! A good book on this topic is How to Lie with Statistics, by Darrell Huff.
In order to calculate a percent, you first have to calculate a proportion, which is a fraction, usually expressed as a decimal, comparing a small number of cases to a total number of cases.
excludes missing values in the computation. A missing value, for example, occurs when an individual does not answer a question in a survey and is therefore not included in the distribution for that question (the "N" is reduced). If there are no missing values, percent and valid percent are the same.
Cumulative Frequency How-To
successive additions or "accumulations" of entries from the frequency (f) column. How to do this from the 4th column starting at the bottom of the frequency column: The first entry in f (bottom) is 1, so enter 1 in cumulative frequency. The second entry in f is 3, so add 3 to 1 to get 4 in CF The third entry in f is 2, so add 2 to 4 to get 6 in CF And so on until the top number in your CF column should equal your N
shows the percentage of cases above or below each score. can be obtained by the following formula: C% = CF/N x 100
Percentage Rate of Change
measures changing conditions over time and can show either a percentage increase or decrease. Example: Historical study of the change in a library's collection size: 1965—collection size 50,000 1975—collection size 150, 000 Rate of change = Later Value - Earlier Value Earlier Value 150,000 - 50,000 = 2 50,000 2 X 100 changes to 200% The library's collection size increased by 200%
the number of cases with property 'x' compared to the number of cases with property 'y; Property X—on a given day, a library checked out 150 fiction titles. Property Y—on the same day, a library checked out 300 nonfiction titles... fiction/nonfiction is150/300 This fraction can be reduced by dividing both sides of the fraction by 150. So, it becomes ½ or 1:2—fiction is ½ of nonfiction. SPSS changes this to a decimal value, .5.
Measures of Central Tendency
Mean, Median and Mode - a number that is located at the center of a group of scores. Keywords include, center value, average value, and most typical value.
the value of the variable that occurs most frequently, A distribution can have more than one
a point in a distribution where half of the cases are above and half of the cases are below when your values are ordered highest to lowest (or lowest to highest).
Median (important note re:)
with a frequency distribution: The median is at the midpoint of all of the scores
What you think of as the "average"
Measures of Central Tendency (important note)
The mode and median are not influenced by extreme (high or low) values in a distribution. The mean is most definitely influenced by extremes.
Skew - what to do about it
In cases where you have an extreme value (high or low) in a distribution, it is important to report both the median and the mean. A very high or low score in the distribution could throw off the mean in the direction of the extreme. often occurs in distributions like income and salary.
STEM AND LEAF DISPLAY
A method of organizing numerical data in which the "stem" values (leading digits of the obervation) are listed in a column and the "leaf" (trailing digits for each observation is listed beside the corresponding stem
average distance from the average
describes how flat or how peaked a curve is
A curve or distribution of scores that has a lot of variance
Distributions that are high and thin.
about halfway between platykurtic and leptokurtic extremes-perfectly symmetrical
a normally distributed set of sepcially scaled values whose mean is always equal to zero and SD must equal 1
a converted z score with the mean always set at 50 and SD 10
characteristics of a normal distribution
1. unimodal and symmetric 2. mesokurtic 3. asymptotic to the xaxis 4. inflection point +- 1 SD
grade equivalent scores
based on relating a given students score on a test to the average scores found at the same time of year for other students roughly the same age and in a particular grade
stem and leaf display
presents only the first digit in the stem column while in the leaf column we find the trailing digits