Chapter 12 - Understanding investment data Flashcards
populations
samples
parameters
- populations - all members of a specified group
- samples - subset of the population that provides info on the population
- parameters - used to describe a charetristic of a population. for example, an average or percentage
normal distribution
- commanuly used for asset returns
- 68.3% of observations lie between 1 standard deviation
- 95.5% of observations lie between 2 standard devations
- 99.75% of observations lie between 3 standard devations
linear regression
- measure of the linear relationshio between two variables - regression looks at the relationship between a dependent variable and one or more independent variables
- if a linear relationship can be established, predection of one variable can be made when the other is known
correlation
r = 1
r= 0
r= -1
measures the linear relationship between two variables
r=1 - returns move together realtive to their mean returns
r= 0 - there is no linear relationship bnetween the variables
r=-1 - returns move in the opposite direction relative to their mean returns
frequency distributions
- grouping of data into a number of non-overlapping classes or intervals. the number of each internval can then be counted and the data can be analysed
- the number of observations in each interval is called the absolute frequency
- the relative frequency is the percentage of the total observations falling into the interval
graphs and absolute frequency
- histogram is a bar chart showing the absolute frequency on the vertical axis and the intervals on the horizontal axis
- frequency polygons is a trend line showing the absolute freqency on the vertical axis and and the interval midpoints on the horizontal axis
sampling methods
sampling bias
methods
- simple random sample
- systematic random sampling - every nth item
- stratified random sampling - divide the population into subgroups (strata) and select a sample from each group
sampling bias
- data-mining bias - models are derived from searching through historic data for patterns or trading rules
- selection bias - data is excluded from the analysis, possibly because it was not available
- look-ahead bias - when a test uses information that was not available at the test date
- survivorship bias - occurs when companies have gone bankrupt, or funds or portfolios that have not been liquidated, are not included in the analysis
other biases
- time-period bias - the test period does not match the conclusion drawn
- data-snooping bias - results of other analysts research are used, or the focus is on patterns that may have been identified by other researchers
geometric mean and arithmetic mean
geometric mean - most suitable for calculating average historic returns, since it compounds the returns and represents what an investor would actually have achieved
arithmetic mean - is often used for projecting future returns since it better predicts the future portfolio value