Sentences

Analysis file

Analysis file

Analysis file

Research questions

Main research question:

How do news media report about cancer screening in the Netherlands from 2010 to 2022?

Subquestions:

  1. What are the volume, characteristics and content of news media reports about cancer screening?
  2. To what extent and how does the media coverage change over time?
  3. How is information about cancer screening in news media reports contextualized?
  4. To what extent and how does the media coverage differ between types of cancer screening?

Q1: Volume, characteristics and content

Volume (N and %)

NB: the number of screening articles is similar to the entire dataset that will be analyzed below (duplicates included). For the articles about cancer, we took the “cancer” part of the search string and registered the number of articles about cancer per year. The percentage reflects the total amount of articles about screening divided by the total number of news articles about the specific cancer type.

Total amount of news articles
News about screening and cancer from 2010 to 2022
cancer_type screening cancer percentage
cervical 1991 8637 23.05
breast 5227 45528 11.48
colon 2895 18946 15.28
prostate 1079 14897 7.24
lung 796 24215 3.29

Characteristics (N and %)

Article characteristics

Inclusion

inclusion n perc
Exclusie 249 26.95
Inclusie 675 73.05

Average word count

Total
m sd
812.48 6076.42
Inclusion vs. exclusion

No significant difference in the number of words for in and excluded articles.

m sd
Exclusie 1606.61 11640.43
Inclusie 519.53 622.23 
##
##  Welch Two Sample t-test
##
## data:  df$word_count by df$inclusion
## t = 1.4729, df = 248.52, p-value = 0.142
## alternative hypothesis: true difference in means between group Exclusie and group Inclusie is not equal to 0
## 95 percent confidence interval:
##  -366.5844 2540.7585
## sample estimates:
## mean in group Exclusie mean in group Inclusie
##              1606.6145               519.5274

Geographical reach

n perc
regional 279 41.33
national 218 32.30
local 178 26.37

News genre

n perc
Nieuwsbericht 441 65.33
Interview en persoonlijk verhaal 146 21.63
Columns en opiniestukken 47 6.96
Achtergrondartikel 23 3.41
Ingezonden brieven 9 1.33
Overig 9 1.33

News content

n perc
Gezondheid en zorg 314 46.52
Beleid 93 13.78
Wetenschap 88 13.04
Cultuur, media, sport en entertainment 71 10.52
COVID-19 52 7.70
Technologie en economie 33 4.89
Overig 24 3.56

Content characteristics

General

Cancer type
n perc
Borstkanker 329 48.74
Darmkanker 150 22.22
Baarmoederhalskanker 82 12.15
Algemeen 76 11.26
Prostaatkanker 28 4.15
Longkanker 10 1.48
Main or side issue
n perc
Bijzaak 172 25.48
Hoofdzaak 503 74.52
Cancer topic
n perc
Kanker 172 25.48
Kankerscreening: bestaande methoden 464 68.74
Kankerscreening: innovatieve methoden 39 5.78
Cancer topic & main/side issue
Bijzaak Hoofdzaak
n % n %
Kanker 42 24.42 130 25.84
Kankerscreening: bestaande methoden 128 74.42 336 66.80
Kankerscreening: innovatieve methoden 2 1.16 37 7.36
Preventive measures
n perc
Screening alternatives 53 7.85
Lifestyle 52 7.70
Vaccination 17 2.52
Surgery 17 2.52
Sources
n perc
Health care professional 175 25.93
Science 160 23.70
Health organization 159 23.56
Personal 142 21.04
Politics 69 10.22
Other 61 9.04

Framing

Effectiveness frame
All
n perc
Expliciet effectief 221 32.74
Impliciet effectief 175 25.93
Neutraal 153 22.67
Effectiviteit betwist 126 18.67
Current

Lung and prostate excluded

n perc
Expliciet effectief 191 34.05
Neutraal 141 25.13
Impliciet effectief 137 24.42
Effectiviteit betwist 92 16.40
Risk frame
All
n perc
Neutral 538 79.70
High 115 17.04
Low 22 3.26
Current

Lung and prostate excluded

n perc
Neutral 460 82.00
High 82 14.62
Low 19 3.39
Consequence frame
All
n perc
Neutral 340 50.37
Severe 320 47.41
Not severe 15 2.22
Current

Lung and prostate excluded

n perc
Neutral 300 53.48
Severe 251 44.74
Not severe 10 1.78
Pros and cons
n perc
Current: pro 467 69.19
Current: con 433 64.15
Innovation: pro 146 21.63
Innovation: con 136 20.15

Text

Q2: Changes over time

Volume of news over time

All articles

NB: rolling mean set at 3, i.e. taking into account one month/quarter before and after the point of measurement.

Per quarter

Per month

Cervical cancer

Per quarter

Per month - dupl incl.

Per month - single docs

Vs. cancer articles

Breast cancer

Per quarter

Per month - dupl incl

Per month - single docs

Vs. cancer articles

Colon cancer

Per quarter

Per month - dupl incl

Per month - single docs

Vs. cancer articles

Lung cancer

Per quarter

Per month - dupl incl

Per month - single docs

Vs. cancer articles

Prostate cancer

Per quarter

Per month - dupl incl

Per month - single docs

Vs. cancer articles

News framing over time

All graphs display a rolling average of 3 (i.e. the average of one quarters/years preceding and following the data point)

Effectiveness (all subcats)

Per year

Per quarter

Per quarter - effectivess disputed

There seems to be a peek in the number of news articles that dispute the effectiveness of cancer screening surrounding the introduction of the colon cancer screening program. To check whether the peek in effectivenss disputed may be related to articles about colon cancer screening, the plot below demonstrates only articles where that were coded as ‘effectiveness disputed’, differentiated between breast, cervical and colon cancer screening.

Effectiveness disputed plot only for colon cancer Including absolute frequencies

Effectiveness (merged)

Per year

Per quarter

Risk framing

Per year

Per quarter

Consequences framing

Per year

Per quarter

Pros and cons (binary)

Per year

Per quarter

News genre

Per year

Per quarter

Q3: Contextualization

Sources

Which sources are quoted in news articles about cancer screening? And how does this relate to the extent to which screening effectiveness and cancer risks are discussed? E.g. is screening described as effective more frequently and cancer risks as higher for articles that quote individuals who are personally involved, whereas articles that quote healthcare professionals more frequently refer to the debate surrounding the effectiveness of screening?

Sources - separate

If a source is present (vs. absent), then what is the likelihood that screening effectiveness, cancer risks and consequences are discussed (and in what direction).

Six groups of sources were coded:

  • Health care professionals
  • Science
  • Health organizations
  • Personal
  • Politics
  • Other

First, we display the absolute and relative frequency tables of each main coding category (effectiveness, risk frame, consequence frame). The percentages represent whether a source is PRESENT (vs. absent) for various subcategories. E.g. for articles in which the effectiveness of screening is disputed/‘betwist’, 55 articles (44%) cite a health professional, and in 66 articles (52%) science is used as a source.

Effectiveness

HCP Science Organization Personal Politics Other
n % n % n % n % n % n %
Effectiviteit betwist 55 31.43 66 41.25 31 19.50 24 16.90 12 17.39 14 22.95
Expliciet effectief 53 30.29 42 26.25 52 32.70 45 31.69 18 26.09 8 13.11
Impliciet effectief 48 27.43 34 21.25 58 36.48 56 39.44 26 37.68 23 37.70
Neutraal 19 10.86 18 11.25 18 11.32 17 11.97 13 18.84 16 26.23

Effectiveness - merged

HCP Science Organization Personal Politics Other
n % n % n % n % n % n %
Effectiviteit betwist 55 31.43 66 41.25 31 19.50 24 16.90 12 17.39 14 22.95
Effectief 101 57.71 76 47.50 110 69.18 101 71.13 44 63.77 31 50.82
Neutraal 19 10.86 18 11.25 18 11.32 17 11.97 13 18.84 16 26.23

Risk

HCP Science Organization Personal Politics Other
n % n % n % n % n % n %
Neutral 124 70.86 119 74.38 110 69.18 95 66.90 56 81.16 50 81.97
High 41 23.43 32 20.00 44 27.67 39 27.46 13 18.84 10 16.39
Low 10 5.71 9 5.62 5 3.14 8 5.63 NA NA 1 1.64

Consequences

HCP Science Organization Personal Politics Other
n % n % n % n % n % n %
Neutral 71 40.57 77 48.12 66 41.51 23 16.20 43 62.32 31 50.82
Severe 99 56.57 80 50.00 90 56.60 109 76.76 23 33.33 25 40.98
Not severe 5 2.86 3 1.87 3 1.89 10 7.04 3 4.35 5 8.20

Sources - combined

Next, we compare whether the main categories (effectiveness, risk, consequences) are distributed differently for news articles in which differnet kinds of sources appear. Two professional sources are grouped as professional (i.e. healthcare professionsals, science) and compared to sources who share stories/information based on personal experiences.

(left out: politics, organization, and ‘other’)

  • Professional = health care professional, science
  • Personal = personal experience
  • Both = professionals and personal experience

Effectiveness

Freqs
effectiveness Both None Personal Professional
n % n % n % n %
Effectiviteit betwist 11 29.73 26 8.41 13 12.38 76 33.93
Expliciet effectief 11 29.73 108 34.95 34 32.38 68 30.36
Impliciet effectief 10 27.03 65 21.04 46 43.81 54 24.11
Neutraal 5 13.51 110 35.60 12 11.43 26 11.61
Chi
##
##  Pearson's Chi-squared test
##
## data:  df_incl$effectiveness and df_incl$source_combi
## X-squared = 109.17, df = 9, p-value < 2.2e-16
##                        df_incl$source_combi
## df_incl$effectiveness          Both        None    Personal Professional
##   Effectiviteit betwist  1.55755311 -4.17131077 -1.49078804   5.28687889
##   Expliciet effectief   -0.32008768  0.67915337 -0.06443139  -0.62346596
##   Impliciet effectief    0.13154093 -1.68830156  3.59900465  -0.53461086
##   Neutraal              -1.16943951  4.77476950 -2.41876418  -3.47669512

Effectiveness - merged

Freqs
effectiveness_merged Both None Personal Professional
n % n % n % n %
Effectiviteit betwist 11 29.73 26 8.41 13 12.38 76 33.93
Effectief 21 56.76 173 55.99 80 76.19 122 54.46
Neutraal 5 13.51 110 35.60 12 11.43 26 11.61
Chi
##
##  Pearson's Chi-squared test
##
## data:  df_incl$effectiveness_merged and df_incl$source_combi
## X-squared = 98.675, df = 6, p-value < 2.2e-16
##                             df_incl$source_combi
## df_incl$effectiveness_merged       Both       None   Personal Professional
##        Effectiviteit betwist  1.5575531 -4.1713108 -1.4907880    5.2868789
##        Effectief             -0.1516764 -0.6149721  2.3443771   -0.8211523
##        Neutraal              -1.1694395  4.7747695 -2.4187642   -3.4766951

Risk

Freqs
cancer_risk Both None Personal Professional
n % n % n % n %
Neutral 18 48.65 273 88.35 77 73.33 170 75.89
High 15 40.54 33 10.68 24 22.86 43 19.20
Low 4 10.81 3 0.97 4 3.81 11 4.91
Chi

Fisher’s exact with ‘low’

##
##  Fisher's Exact Test for Count Data
##
## data:  df_incl$cancer_risk and df_incl$source_combi
## p-value = 6.718e-08
## alternative hypothesis: two.sided
##                    df_incl$source_combi
## df_incl$cancer_risk       Both       None   Personal Professional
##             Neutral -2.1158940  1.7023377 -0.7311729   -0.6388608
##             High     3.4636694 -2.7074689  1.4448691    0.7829939
##             Low      2.5443544 -2.2281733  0.3123250    1.3690878

Chi with ‘low’ excluded

##
##  Pearson's Chi-squared test
##
## data:  df_q4_risk$cancer_risk and df_q4_risk$source_combi
## X-squared = 31.07, df = 3, p-value = 8.219e-07
##                       df_q4_risk$source_combi
## df_q4_risk$cancer_risk       Both       None   Personal Professional
##                Neutral -1.7621649  1.3156421 -0.6810776   -0.4143148
##                High     3.8114372 -2.8456402  1.4731223    0.8961334

Consequences

Freqs
cancer_consequences Both None Personal Professional
n % n % n % n %
Neutral 7 18.92 208 67.31 16 15.24 109 48.66
Severe 28 75.68 100 32.36 81 77.14 111 49.55
Not severe 2 5.41 1 0.32 8 7.62 4 1.79
Chi

Fisher’s exact

##
##  Fisher's Exact Test for Count Data
##
## data:  df_incl$cancer_consequences and df_incl$source_combi
## p-value < 2.2e-16
## alternative hypothesis: two.sided
##                            df_incl$source_combi
## df_incl$cancer_consequences       Both       None   Personal Professional
##                  Neutral    -2.6955909  4.1965844 -5.0723984   -0.3605331
##                  Severe      2.4973359 -3.8410223  4.4253340    0.4665131
##                  Not severe  1.2988792 -2.2388161  3.7097041   -0.4382505

Chi with ‘not severe’ excluded

##
##  Pearson's Chi-squared test
##
## data:  df_q4_cons$cancer_consequences and df_q4_cons$source_combi
## X-squared = 93.525, df = 3, p-value < 2.2e-16
##                               df_q4_cons$source_combi
## df_q4_cons$cancer_consequences       Both       None   Personal Professional
##                        Neutral -2.5976817  3.9164953 -4.8054970   -0.4070458
##                        Severe   2.6776290 -4.0370310  4.9533930    0.4195732

Type of news article

The fact that we find differences for when both or none appear as sources in a news article, may be related to the fact that these reflect a different kind of journalistic genre (i.e. news article vs. background or interview). Therefore: extra test for genre X frames (efffectiveness, risk, consequences).

Freqs
effectiveness Interview Column News
n % n % n %
Effectiviteit betwist 14 29.79 26 17.81 71 16.10
Effectief 21 44.68 99 67.81 259 58.73
Neutraal 12 25.53 21 14.38 111 25.17
Risk Interview Column News
n % n % n %
Neutral 33 70.21 108 73.97 365 82.77
High 11 23.40 30 20.55 67 15.19
Low 3 6.38 8 5.48 9 2.04
Consequences Interview Column News
n % n % n %
Neutral 20 42.55 36 24.66 266 60.32
Severe 27 57.45 106 72.60 166 37.64
Not severe NA NA 4 2.74 9 2.04
Chi
##
##  Pearson's Chi-squared test
##
## data:  df_q4_extra$effectiveness_merged and df_q4_extra$newsgenre
## X-squared = 13.799, df = 4, p-value = 0.007965
##                        df_q4_extra$newsgenre
##                         Columns en opiniestukken
##   Effectiviteit betwist               2.01190450
##   Effectief                          -1.33876032
##   Neutraal                            0.40551272
##                        df_q4_extra$newsgenre
##                         Interview en persoonlijk verhaal Nieuwsbericht
##   Effectiviteit betwist                       0.08672858   -0.70670799
##   Effectief                                   1.25477264   -0.28492367
##   Neutraal                                   -2.11179655    1.08270855
##
##  Pearson's Chi-squared test
##
## data:  df_q4_extra$effectiveness_merged and df_q4_extra$newsgenre
## X-squared = 13.799, df = 4, p-value = 0.007965
##                        df_q4_extra$newsgenre
##                         Columns en opiniestukken
##   Effectiviteit betwist               2.01190450
##   Effectief                          -1.33876032
##   Neutraal                            0.40551272
##                        df_q4_extra$newsgenre
##                         Interview en persoonlijk verhaal Nieuwsbericht
##   Effectiviteit betwist                       0.08672858   -0.70670799
##   Effectief                                   1.25477264   -0.28492367
##   Neutraal                                   -2.11179655    1.08270855
##
##  Pearson's Chi-squared test
##
## data:  df_q4_extra_cons$cancer_consequences and df_q4_extra_cons$newsgenre
## X-squared = 57.925, df = 2, p-value = 2.64e-13
##          df_q4_extra_cons$newsgenre
##           Columns en opiniestukken Interview en persoonlijk verhaal
##   Neutral               -0.8852933                       -4.3853437
##   Severe                 0.9187123                        4.5508862
##          df_q4_extra_cons$newsgenre
##           Nieuwsbericht
##   Neutral     2.8062430
##   Severe     -2.9121760

Risk and effectiveness frames

Risk vs. consequence

Do news articles that mention the risks of cancer also refer to the severity of the consequences?

Freqs

Risk
Consequences Neutral High Low
n % n % n %
Neutral 317 58.92 17 14.78 6 27.27
Severe 210 39.03 94 81.74 16 72.73
Not severe 11 2.04 4 3.48 NA NA

Chi-square

“low” risk and consequence excluded

##
##  Pearson's Chi-squared test with Yates' continuity correction
##
## data:  df_q4_consrisk$cancer_risk and df_q4_consrisk$cancer_consequences
## X-squared = 72.105, df = 1, p-value < 2.2e-16
##
## df_q4_consrisk$cancer_risk   Neutral    Severe
##                    Neutral  2.475006 -2.594255
##                    High    -5.392872  5.652708

Screening effectiveness vs. risk

When cancer risks are mentioned, is screening presented as an effective method to reduce cancer risks?

Effectiveness

Freqs
Effectiveness
Risk Betwist Effective:expl Effectiv:impl Neutral
n % n % n % n %
Neutral 86 68.25 161 72.85 144 82.29 147 96.08
High 31 24.60 53 23.98 25 14.29 6 3.92
Low 9 7.14 7 3.17 6 3.43 NA NA
Chi-square

‘Low’ risk excluded

##
##  Pearson's Chi-squared test
##
## data:  df_q4_risk$effectiveness and df_q4_risk$cancer_risk
## X-squared = 34.603, df = 3, p-value = 1.478e-07
##                         df_q4_risk$cancer_risk
## df_q4_risk$effectiveness    Neutral       High
##    Effectiviteit betwist -1.0587689  2.2900417
##    Expliciet effectief   -1.1531929  2.4942741
##    Impliciet effectief    0.4036169 -0.8729945
##    Neutraal               1.8655092 -4.0349636

Effectiveness - merged

Freqs
Effectiveness_merged
Risk Betwist Effectief Neutral
n % n % n %
Neutral 86 68.25 305 77.02 147 96.08
High 31 24.60 78 19.70 6 3.92
Low 9 7.14 13 3.28 NA NA
Chi-square

‘Low’ risk excluded

##
##  Pearson's Chi-squared test
##
## data:  df_q4_risk$effectiveness_merged and df_q4_risk$cancer_risk
## X-squared = 28.129, df = 2, p-value = 7.795e-07
##                        df_q4_risk$cancer_risk
##                            Neutral       High
##   Effectiviteit betwist -1.0587689  2.2900417
##   Effectief             -0.5938942  1.2845508
##   Neutraal               1.8655092 -4.0349636

Effectiveness vs. consequences

Effectiveness

Freqs
Effectiveness
Consequences Betwist Effective:expl Effectiv:impl Neutral
n % n % n % n %
Neutral 47 37.30 67 30.32 92 52.57 134 87.58
Severe 75 59.52 149 67.42 78 44.57 18 11.76
Not severe 4 3.17 5 2.26 5 2.86 1 0.65
Chi-square

‘Not severe’ consequence excluded

##
##  Pearson's Chi-squared test
##
## data:  df_q4_cons$effectiveness and df_q4_cons$cancer_consequences
## X-squared = 126.75, df = 3, p-value < 2.2e-16
##                         df_q4_cons$cancer_consequences
## df_q4_cons$effectiveness    Neutral     Severe
##    Effectiviteit betwist -1.9991268  2.0606527
##    Expliciet effectief   -4.1970284  4.3261979
##    Impliciet effectief    0.4727668 -0.4873169
##    Neutraal               6.2942253 -6.4879389

Effectiveness - merged

Freqs
Effectiveness
Consequences Betwist Effectief Neutral
n % n % n %
Neutral 47 37.30 159 40.15 134 87.58
Severe 75 59.52 227 57.32 18 11.76
Not severe 4 3.17 10 2.53 1 0.65
Chi-square

‘Not severe’ consequence excluded

##
##  Pearson's Chi-squared test
##
## data:  df_q4_cons$effectiveness_merged and df_q4_cons$cancer_consequences
## X-squared = 106.42, df = 2, p-value < 2.2e-16
##                        df_q4_cons$cancer_consequences
##                           Neutral    Severe
##   Effectiviteit betwist -1.999127  2.060653
##   Effectief             -2.825860  2.912830
##   Neutraal               6.294225 -6.487939

Pros and cons

Presence or absence per news article

Pro vs con - current

Freqs

Pros
Cons Absent Present
n % n %
Absent 105 50.48 137 29.34
Present 103 49.52 330 70.66

Chi-square

##
##  Pearson's Chi-squared test with Yates' continuity correction
##
## data:  df_incl$current_neg_bin and df_incl$current_pos_bin
## X-squared = 27.064, df = 1, p-value = 1.969e-07
##                        df_incl$current_pos_bin
## df_incl$current_neg_bin    Absent   Present
##                 Absent   3.523612 -2.351588
##                 Present -2.634218  1.758025

Pro vs. con - innovative

Freqs

Pros
Cons Absent Present
n % n %
Absent 526 99.43 13 8.9
Present 3 0.57 133 91.1

Chi-square

##
##  Fisher's Exact Test for Count Data
##
## data:  df_incl$innov_neg_bin and df_incl$innov_pos_bin
## p-value < 2.2e-16
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##   471.6894 8192.0000
## sample estimates:
## odds ratio
##   1716.035
##                      df_incl$innov_pos_bin
## df_incl$innov_neg_bin     Absent    Present
##               Absent    5.039891  -9.593399
##               Present -10.033352  19.098420

Current vs. innovative

Pro, con, both, absent (4 levels)

Freqs - all

Current measures
Innovative measures Absent Both Con Pro
n % n % n % n %
Absent 99 94.29 214 64.85 80 77.67 133 97.08
Both 4 3.81 113 34.24 14 13.59 2 1.46
Con 2 1.90 NA NA 1 0.97 NA NA
Pro NA NA 3 0.91 8 7.77 2 1.46

Freqs - binary

Since oftentimes, pro and cons are given for current measures, this is changed to “procon” present vs. absent.

Current measures
Innovative measures Absent Both Con Pro
n % n % n % n %
Absent 99 94.29 214 64.85 80 77.67 133 97.08
Present 6 5.71 116 35.15 23 22.33 4 2.92

Chi-square

##
##  Fisher's Exact Test for Count Data
##
## data:  df_incl$innov_procon_bin and df_incl$current_procon
## p-value < 2.2e-16
## alternative hypothesis: two.sided
##                         df_incl$current_procon
## df_incl$innov_procon_bin      Absent        Both         Con         Pro
##                  Absent   1.89902790 -2.69115527 -0.02943450  2.53972726
##                  Present -3.56805124  5.05636590  0.05530398 -4.77185038

Effectiveness vs. current pro/con

Pro, con, both, absent (4 levels)

Freqs

Current pros and cons
effectiveness Absent Both Con Pro
n % n % n % n %
Effectiviteit betwist 2 1.90 84 25.45 40 38.83 NA NA
Expliciet effectief 1 0.95 130 39.39 5 4.85 85 62.04
Impliciet effectief 42 40.00 72 21.82 27 26.21 34 24.82
Neutraal 60 57.14 44 13.33 31 30.10 18 13.14

Chi-square

De tabel geeft met name inzicht hoe de coderingen zijn gedaan. Maar gezien er inhoudelijke overlap zit tussen de ‘effectiveness’ codering in het algemeen en de pros and cons op zinsniveau, heeft het geen zin om daar een chikwadraat over te berekenen.

Q4: Differences between screening

NB: Chi-squares only between breast, cervical, and colon cancer. Frequencies for lung cancer and prostate cancer are too low.

Framing

Effectiveness

Freqs

Cervix-breast-colon
Cervical Breast Colon
n % n % n %
Effectiviteit betwist 8 9.76 52 15.81 32 21.33
Expliciet effectief 29 35.37 110 33.43 52 34.67
Impliciet effectief 30 36.59 70 21.28 37 24.67
Neutraal 15 18.29 97 29.48 29 19.33
Lung-psa
Lung Prostate
n % n %
Effectiviteit betwist 1 10 18 64.29
Expliciet effectief 3 30 5 17.86
Impliciet effectief 5 50 3 10.71
Neutraal 1 10 2 7.14

Chisquare

##
##  Pearson's Chi-squared test
##
## data:  df_q3$effectiveness and df_q3$cancer_type
## X-squared = 16.915, df = 6, p-value = 0.009601
##                        df_q3$cancer_type
## df_q3$effectiveness     Baarmoederhalskanker Borstkanker  Darmkanker
##   Effectiviteit betwist          -1.48549461 -0.26597282  1.49223208
##   Expliciet effectief             0.20477817 -0.19015068  0.13020473
##   Impliciet effectief             2.22909750 -1.15401886  0.06096534
##   Neutraal                       -1.23565955  1.57368726 -1.41700764

Effectiveness - effective merged

Freqs

Cervix-breast-colon
Cervical Breast Colon
n % n % n %
Effectiviteit betwist 8 9.76 52 15.81 32 21.33
Effectief 59 71.95 180 54.71 89 59.33
Neutraal 15 18.29 97 29.48 29 19.33
Lung-psa
Lung Prostate
n % n %
Effectiviteit betwist 1 10 18 64.29
Effectief 8 80 8 28.57
Neutraal 1 10 2 7.14

Chisquare

##
##  Pearson's Chi-squared test
##
## data:  df_q3$effectiveness_merged and df_q3$cancer_type
## X-squared = 13.879, df = 4, p-value = 0.007693
##                           df_q3$cancer_type
## df_q3$effectiveness_merged Baarmoederhalskanker Borstkanker Darmkanker
##      Effectiviteit betwist           -1.4854946  -0.2659728  1.4922321
##      Effectief                        1.5968955  -0.8909273  0.1387598
##      Neutraal                        -1.2356595   1.5736873 -1.4170076

Risk frame

Freqs

Cervix-breast-colon
Cervical Breast Colon
n % n % n %
Neutral 64 78.05 281 85.41 115 76.67
High 11 13.41 42 12.77 29 19.33
Low 7 8.54 6 1.82 6 4.00
Lung-psa
Lung Prostate
n % n %
Neutral 3 30 17 60.71
High 7 70 10 35.71
Low NA NA 1 3.57

Chi-square

##
##  Pearson's Chi-squared test
##
## data:  df_q3$cancer_risk and df_q3$cancer_type
## X-squared = 13.239, df = 4, p-value = 0.01016
##                  df_q3$cancer_type
## df_q3$cancer_risk Baarmoederhalskanker Borstkanker Darmkanker
##           Neutral           -0.3947740   0.6838348 -0.7208692
##           High              -0.2847278  -0.8780749  1.5109396
##           Low                2.5339609  -1.5405991  0.4080805

Consequences frame

Freqs

Cervix-breast-colon
Cervical Breast Colon
n % n % n %
Neutral 44 53.66 175 53.19 81 54.00
Severe 38 46.34 148 44.98 65 43.33
Not severe NA NA 6 1.82 4 2.67
Lung-psa
Lung Prostate
n % n %
Neutral 1 10 9 32.14
Severe 9 90 16 57.14
Not severe NA NA 3 10.71

Chi-square

Chi-square not reliable due to low frequencies. Therefore: fisher’s exact.

And chi-square with “low” excluded.

##
##  Fisher's Exact Test for Count Data
##
## data:  df_q3$cancer_consequences and df_q3$cancer_type
## p-value = 0.7688
## alternative hypothesis: two.sided
##                          df_q3$cancer_type
## df_q3$cancer_consequences Baarmoederhalskanker Borstkanker  Darmkanker
##                Neutral              0.02261155 -0.07055362  0.08777097
##                Severe               0.21659708  0.06596751 -0.25784246
##                Not severe          -1.20899776  0.05594151  0.81104673
##
##  Pearson's Chi-squared test
##
## data:  df_q3_2$cancer_consequences and df_q3_2$cancer_type
## X-squared = 0.092616, df = 2, p-value = 0.9547
##                            df_q3_2$cancer_type
## df_q3_2$cancer_consequences Baarmoederhalskanker Borstkanker  Darmkanker
##                     Neutral          -0.09669558 -0.06500637  0.16915630
##                     Severe            0.10571349  0.07106892 -0.18493194

Pros and cons: number of arguments

Current methods

Freqs

Cervical Breast Colon
n % n % n %
Absent 9 10.98 57 17.33 26 17.33
Both 50 60.98 141 42.86 81 54.00
Con 13 15.85 49 14.89 15 10.00
Pro 10 12.20 82 24.92 28 18.67

Chisquare

##
##  Pearson's Chi-squared test
##
## data:  df_q3$current_procon and df_q3$cancer_type
## X-squared = 15.248, df = 6, p-value = 0.01841
##                     df_q3$cancer_type
## df_q3$current_procon Baarmoederhalskanker Borstkanker Darmkanker
##               Absent           -1.2127975   0.4147332  0.2824890
##               Both              1.6243994  -1.4659741  0.9700633
##               Con               0.5201744   0.5719052 -1.2315871
##               Pro              -1.8003684   1.3858327 -0.7212684

Innovative methods

Freqs

Cervical Breast Colon
n % n % n %
Absent 67 81.71 259 78.72 115 76.67
Both 12 14.63 65 19.76 30 20.00
Con 1 1.22 1 0.30 1 0.67
Pro 2 2.44 4 1.22 4 2.67

Chisquare

Fisher’s exact due to low frequencies.

##
##  Fisher's Exact Test for Count Data
##
## data:  df_q3$innov_procon and df_q3$cancer_type
## p-value = 0.5118
## alternative hypothesis: two.sided

Pros and cons: content (post-hoc analysis)

Not done for now, first qualitative post hoc analysis needed, and then assess whether these qualitative themes can be quantified.