Topic 1: Statistical analysis Assessment statement

State that error bars are a graphical representation of the variability of data.

IBO Notes
Error bars can be used to show either the range of the data or the standard deviation. Students should specify the standard deviation (s), not the population standard deviation. Students will not be e pected to !now the formulas for calculating these statistics. "hey will be e pected to use the standard deviation function of a graphic display or scientific calculator. #im $% Students could also be taught how to calculate standard deviation using a spreadsheet computer program.

Student Notes

1.1.1

1.1.2

Calculate the mean and standard deviation of a set of values.

1.1.&

State that the term standard deviation is used to summari'e the spread of values around the mean, and that ()* of the values fall within one standard deviation of the mean. E plain how the standard deviation is useful for comparing the means and the spread of data between two or more samples.

+or normally distributed data, about ()* of all values lie within ,1 standard deviation (s or -) of the mean. "his rises to about ./* for ,2 standard deviations.

1.1.0

# small standard deviation indicates that the data is clustered closely around the mean value. Conversely, a large standard deviation indicates a wider spread around the mean. +or the t2test to be applied, the data must have a normal distribution and a sample si'e of at least 13. "he t2test can be used to compare two sets of data and measure the amount of overlap. Students will not be e pected to calculate values of t. 4nly a two2tailed, unpaired t2test is e pected. #im $% 5hile students are not e pected to calculate a value for the t2test, students could be shown how to calculate such values using a spreadsheet program or the graphic display calculator. TOK: The scientific community defines an objective standard by which claims about data can be made. #im $% 5hile calculations of such values are not e pected, students who want to use r and r2 values in their practical wor! could be shown how to determine such values using a spreadsheet program.

1.1./

1educe the significance of the difference between two sets of data using calculated values for t and the appropriate tables.

1.1.(

E plain that the e istence of a correlation does not establish that there is a causal relationship between two variables.

STATISTICS!!!
The science of data

What is data?
Information, in the form of facts or figures obtained from experiments or surveys, used as a basis for making ca cu ations or dra!ing conc usions
"ncarta dictionary

Statistics in Science
# $ata can be co ected about a popu ation %surveys& # $ata can be co ected about a process %experimentation&

' types of $ata

(ua itative (uantitative

(ua itative $ata

# Information that re ates to characteristics or description %observab e )ua ities& # Information is often grouped by descriptive category # "xamp es
* Species of p ant * Type of insect * Shades of co or * +ank of f avor in taste testing Remember: qualitative data can be scored and evaluated numerically

(ua itative data, manipu ated numerica y

# Survey resu ts, teens and need for environmenta action

(uantitative data
# (uantitative * measured using a naturally occurring numerica sca e # "xamp es *Chemica concentration *Temperature *,ength *Weight-etc.

(uantitation
# /easurements are often disp ayed graphica y

(uantitation 0 /easurement
# In data co ection for 1io ogy, data must be measured carefu y, using aboratory e)uipment %ex. Timers, metersticks, p2 meters, ba ances , pipettes, etc& # The imits of the e)uipment used add some uncertainty to the data co ected. A e)uipment has a certain magnitude of uncertainty. 3or examp e, is a ru er that is mass4produced a good measure of 5 cm? 5mm? 6.5mm? # 3or )uantitative testing, you must indicate the level of uncertainty of the tool that you are using for measurement!!

2o! to determine uncertainty?

# 7sua y the instrument manufacturer !i indicate this * read !hat is provided by the manufacturer. # 1e sure that the number of significant digits in the data tab e8graph ref ects the precision of the instrument used %for ex. If the manufacturer states that the accuracy of a ba ance is to 6.5g * and your average mass is '.69g, be sure to round the average to '.5g& :our data must be consistent !ith your measurement too regarding significant figures.

3inding the imits

# As a ;ru e4of4thumb<, if not specified, use =84 58' of the sma est measurement unit %ex metric ru er is ined to 5mm,so the imit of uncertainty of the ru er is =84 6.> mm.& # If the room temperature is read as '> degrees C, !ith a thermometer that is scored at 5 degree interva s * !hat is the range of possib e temperatures for the room? # %ans.s =84 6.> degrees Ce sius 4 if you read 5>oC, it may in fact be 5?.> or 5>.> degrees&

,ooking at $ata
# 2o! accurate is the data? %2o! c ose are the data to the ;rea < resu ts?& This is a so considered as 1IAS # 2o! precise is the data? %A test systems have some uncertainty, due to imits of measurement& "stimation of the imits of the experimenta uncertainty is essentia .

Comparing Averages
# @nce the ' averages are ca cu ated for each set of data, the average va ues can be p otted together on a graph, to visua iAe the re ationship bet!een the '

$ra!ing error bars

# The simp est !ay to dra! an error bar is to use the mean as the centra point, and to use the distance of the measurement that is furthest from the average as the endpoints of the data bar

Ba ue farthest from average Ca cu ated distance

Average va ue

What do error bars suggest?

# If the bars sho! extensive over ap, it is ike y that there is not a significant difference bet!een those va ues

(uick +evie! * C measures of ;Centra Tendency<

# modeD va ue that appears most fre)uent y # medianD When a data are isted from east to greatest, the va ue at !hich ha f of the observations are greater, and ha f are esser. # The most common y used measure of centra tendency is the mean, or arithmetic average %sum of data points divided by the number of points&

2o! can eaf engths be disp ayed graphica y?

Simp y measure the engths of each and p ot ho! many are of each ength

If smoothed, the histogram data assumes this shape

This Shape?
# Is a c assic be 4shaped curve, AEA Faussian $istribution Curve, AEA a Gorma $istribution curve. # "ssentia y it means that in a studies !ith an ade)uate number of datapoints %HC6& a significant number of resu ts tend to be near the mean. 3e!er resu ts are found farther from the mean

# The standard deviation is a statistic that te s you ho! tight y a the various examp es are c ustered around the mean in a set of data

Standard deviation
# The STAG$A+$ $"BIATI@G is a more sophisticated indicator of the precision of a set of a given number of measurements
* The standard deviation is ike an average deviation of measurement va ues from the mean. In arge studies, the standard deviation is used to dra! error bars, instead of the maximum deviation.

A typica standard distribution curve

According to this curveD

# @ne standard deviation a!ay from the mean in either direction on the horiAonta axis %the red area on the preceding graph& accounts for some!here around 9I percent of the data in this group. # T!o standard deviations a!ay from the mean %the red and green areas& account for rough y J> percent of the data.

Three Standard $eviations?

# three standard deviations %the red, green and b ue areas& account for about JJ percent of the data

4Csd 4'sd

=845sd

'sd

=Csd

2o! is Standard $eviation ca cu ated?

With this formu a!

Basic example of Calculating Standard Deviation Consider a population consisting of the following values

"here are eight data points in total, with a mean (or average) value of /%

"o calculate the standard deviation, we compute the difference of each data point from the mean, and s6uare the result%

7e t we average these values and ta!e the s6uare root, which gives the standard deviation%

"herefore, the population above has a standard deviation of 2.

1.

5hat is meant by the following terms8

a.

9ean

b.

Standard deviation

c.

:ange; variability

2.

Error bars can be used to show variability in data (either range of data or standard deviation)

<n which two pairs of groups below can we see an overlap in the standard deviation of the data8 a. b. ========= and ========== ========== and ==========

c. <n which group is the mean 74" li!ely to be significantly different to the mean of group &8

&. 4n the a es below, plot the following three curves (all have a normal distribution)% a. >as the highest mean and a high standard deviation. b. >as the lowest mean, but the highest fre6uency at that mean. c. >as a mean between (a) and (b) and has the smallest standard deviation.

0. Complete the following statements by either filling in the blan! or circling the correct answer; a) , 1- (standard deviation) from the mean represents * of all the data points. b) <n data with a high standard deviation, data are clustered closer to; further from the mean. c) <n data with a low standard deviation, data are clustered closer to; further from the mean. d) 4verlapping standard deviations suggest two datasets are;are not significantly different. * of all data fall within 2 standard deviations of the mean.

Group A Group B 24 24 25 29 26 25 23 23 25 29 25 32 26 34 27 31 23 32 23 29 Mean Stdev P= 0.004 (T-Te t!

(. Calculate the means and standard deviations of these two groups of data (to one decimal place)

Show your wor!ing here.

$. <n a t2test comparing ?roup # and ?roup @, the A value was calculated as 3.330. 5hat does this A value tell us about these two sets of data8

E plain your answer.

). 5hy does the scientific community place so much importance on significance tests such as the t2test8 S I!!S C"#C < can calculate mean and standard deviation using% a. # graphical display or scientific calculator b. 9icrosoft E cel 233$

Boo! at the graph below%

.. 5hich of the following statements are true8 a. "here is no correlation between number of cellphones and number of students with asthma. b. "here is a positive correlation between number of cellphones and asthma cases. c. "here is a negative correlation between number of cellphones and asthma cases d. <ncreased numbers of cellphones cause an increase in numbers of students with asthma. e. <ncreased numbers of asthmatic cause increased numbers of cellphones in a school. f. "here is no evidence of causality in this graph.

13. Suggest two comparisons in which a causal relationship is li!ely (e.g. temperature and rate of reaction).